2004-06-24 14:18:37 +00:00
|
|
|
/* NetworkManager -- Network link manager
|
|
|
|
|
*
|
|
|
|
|
* Dan Williams <dcbw@redhat.com>
|
|
|
|
|
*
|
|
|
|
|
* 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.
|
|
|
|
|
*
|
|
|
|
|
* 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., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
|
|
|
|
|
*
|
|
|
|
|
* (C) Copyright 2004 Red Hat, Inc.
|
|
|
|
|
*/
|
|
|
|
|
|
|
|
|
|
#include <errno.h>
|
|
|
|
|
#include <glib.h>
|
|
|
|
|
#include <dbus/dbus-glib.h>
|
|
|
|
|
#include <hal/libhal.h>
|
|
|
|
|
#include <iwlib.h>
|
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
|
|
|
#include <signal.h>
|
|
|
|
|
#include <string.h>
|
2004-06-24 14:18:37 +00:00
|
|
|
|
|
|
|
|
#include "NetworkManager.h"
|
2004-10-21 17:42:14 +00:00
|
|
|
#include "NetworkManagerMain.h"
|
2004-06-24 14:18:37 +00:00
|
|
|
#include "NetworkManagerDevice.h"
|
|
|
|
|
#include "NetworkManagerUtils.h"
|
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
|
|
|
#include "NetworkManagerDbus.h"
|
|
|
|
|
#include "NetworkManagerWireless.h"
|
|
|
|
|
#include "NetworkManagerPolicy.h"
|
2004-08-02 21:12:40 +00:00
|
|
|
#include "NetworkManagerAPList.h"
|
2004-08-24 00:31:47 +00:00
|
|
|
#include "backends/NetworkManagerSystem.h"
|
2004-06-24 14:18:37 +00:00
|
|
|
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
/* Local static prototypes */
|
2004-06-24 14:18:37 +00:00
|
|
|
static gboolean mii_get_link (NMDevice *dev);
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
static gpointer nm_device_activation_worker (gpointer user_data);
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
/******************************************************/
|
|
|
|
|
|
|
|
|
|
/* Wireless device specific options */
|
|
|
|
|
typedef struct NMDeviceWirelessOptions
|
|
|
|
|
{
|
2004-10-21 17:42:14 +00:00
|
|
|
char *cur_essid;
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
gboolean supports_wireless_scan;
|
2004-08-26 20:05:24 +00:00
|
|
|
guint8 max_quality;
|
2004-09-13 17:43:16 +00:00
|
|
|
guint8 noise;
|
|
|
|
|
gint8 strength;
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
|
2004-08-26 20:05:24 +00:00
|
|
|
GMutex *scan_mutex;
|
2004-10-11 21:32:19 +00:00
|
|
|
/* We keep a couple lists around since wireless cards
|
|
|
|
|
* are a bit flakey and don't report the same access
|
|
|
|
|
* points every time. The lists get merged and diffed
|
|
|
|
|
* to figure out the "real" list, but the latest_ap_list
|
|
|
|
|
* is always the most-current scan.
|
|
|
|
|
*/
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
NMAccessPointList *ap_list;
|
2004-10-11 21:32:19 +00:00
|
|
|
NMAccessPointList *cached_ap_list1;
|
|
|
|
|
NMAccessPointList *cached_ap_list2;
|
|
|
|
|
NMAccessPointList *cached_ap_list3;
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
|
|
|
|
|
NMAccessPoint *best_ap;
|
|
|
|
|
GMutex *best_ap_mutex;
|
|
|
|
|
gboolean freeze_best_ap;
|
|
|
|
|
|
|
|
|
|
gboolean user_key_received;
|
2004-10-07 01:50:39 +00:00
|
|
|
gboolean now_scanning;
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
} NMDeviceWirelessOptions;
|
|
|
|
|
|
|
|
|
|
/* Wired device specific options */
|
|
|
|
|
typedef struct NMDeviceWiredOptions
|
|
|
|
|
{
|
|
|
|
|
int foo;
|
|
|
|
|
} NMDeviceWiredOptions;
|
|
|
|
|
|
|
|
|
|
typedef union NMDeviceOptions
|
|
|
|
|
{
|
|
|
|
|
NMDeviceWirelessOptions wireless;
|
|
|
|
|
NMDeviceWiredOptions wired;
|
|
|
|
|
} NMDeviceOptions;
|
|
|
|
|
|
|
|
|
|
|
2004-09-08 18:14:42 +00:00
|
|
|
typedef struct NMDeviceConfigInfo
|
|
|
|
|
{
|
|
|
|
|
gboolean use_dhcp;
|
|
|
|
|
guint32 ip4_gateway;
|
|
|
|
|
guint32 ip4_address;
|
|
|
|
|
guint32 ip4_netmask;
|
2004-10-15 00:06:14 +00:00
|
|
|
guint32 ip4_broadcast;
|
2004-09-08 18:14:42 +00:00
|
|
|
/* FIXME: ip6 stuff */
|
|
|
|
|
} NMDeviceConfigInfo;
|
|
|
|
|
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
/*
|
|
|
|
|
* NetworkManager device structure
|
|
|
|
|
*/
|
|
|
|
|
struct NMDevice
|
|
|
|
|
{
|
|
|
|
|
guint refcount;
|
|
|
|
|
|
2004-10-21 17:42:14 +00:00
|
|
|
char *udi;
|
|
|
|
|
char *iface;
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
NMDeviceType type;
|
2004-10-21 17:42:14 +00:00
|
|
|
NMDriverSupportLevel driver_support_level;
|
|
|
|
|
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
gboolean link_active;
|
|
|
|
|
guint32 ip4_address;
|
|
|
|
|
/* FIXME: ipv6 address too */
|
|
|
|
|
NMData *app_data;
|
|
|
|
|
NMDeviceOptions options;
|
2004-09-08 18:14:42 +00:00
|
|
|
NMDeviceConfigInfo config_info;
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
|
2004-10-07 01:50:39 +00:00
|
|
|
gboolean activating; /* Set by main thread before beginning activation */
|
|
|
|
|
gboolean just_activated; /* Set by activation thread after successful activation */
|
|
|
|
|
gboolean quit_activation; /* Flag to signal activation thread to stop activating */
|
2004-10-23 06:43:01 +00:00
|
|
|
gboolean activation_failed; /* Did the activation fail? */
|
2004-08-31 Dan Williams <dcbw@redhat.com>
* Remove 'debug' extern global from all files since we now
use syslog()
* src/NetworkManager.[ch]
- Break out routine that get the net.interface property from HAL,
removing that logic from nm_create_device_and_add_to_list()
- (nm_create_device_and_add_to_list): make this a bit more general so
it doesn't do the talking to HAL. Also add arguments to facilitate
the create of test devices.
- (nm_data_mark_state_changed): rename from nm_data_set_state_modified()
- (nm_data_new, main, nm_print_usage): add new argument "--enable-test-devices"
which makes NetworkManager listen for dbus commands to create test
devices, which have no backing hardware. Use when you're on a plane
for example, and/or forgot your wireless card at home. Test devices
_cannot_ be created unless NM is started with --enable-test-devices.
* src/NetworkManagerDbus.[ch]
- New "getLinkActive" method for devices
- New "setLinkActive" method for devices (only works on test devices)
- New "createTestDevice" method on NetworkManager object to create a test
device of a specified type (ie wired, wireless). UDI is created from
scratch, as is the interface name. Only works when NM is started with
--enable-test-devices switch.
- New "removeTestDevice" method on NetworkManager object which removes a
test device. Only works when NM is started with --enable-test-devices
* src/NetworkManagerDevice.[ch]
- Logic to facilitate test devices. Add variables to NMDevice struct to indicate
whether a device is a test device or not, and what its link status is.
- Deal with test devices in most functions. For those that work directly on hardware
special-case test devices.
- (nm_device_new): don't create a test device if test devices weren't enabled on the
command-line.
- (nm_device_update_link_active): split out logic for wired and wireless device link
checking to separate functions to facilitate test device link checking.
- (nm_device_set_enc_key): Since some drivers for wireless cards are daft and
don't make a distinction between System Authentication and Encryption
(namely Cisco aironet), we use Open System auth when setting a WEP key
on the card. We don't deal with Shared Key auth yet.
- (nm_device_activation_worker): split the activation cancel check logic out into
a separate routine nm_device_activation_cancel_if_needed()
- (nm_device_activation_signal_cancel): rename from nm_device_activation_cancel()
- (nm_device_fake_ap_list): Test wireless devices obviously cannot scan, so create
a list of fake access points that they can "see"
- (nm_device_is_test_device): return whether or not a device is a test device
* src/NetworkManagerPolicy.c
- (nm_policy_get_best_device): attempt to deal with wireless network selection,
previously if you "locked"/forced NM to use a wireless device but then
selected a wireless network for NM to use, it would switch to a wired device.
So, if the active device is wireless and it has a "forced" best AP, use it
if the "forced" best AP is still valid
- (nm_state_modification_monitor): deal with NULL best devices, for example
there were no usable network devices, or the last one was removed
* src/backends/NetworkManager*.c
- Deal with test devices, mostly just return success for operations like getting
a DHCP address
* test/nmtestdevices.c
- Test tool to create/remove/link-switch test devices
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@112 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-31 16:09:15 +00:00
|
|
|
|
|
|
|
|
gboolean test_device;
|
|
|
|
|
gboolean test_device_up;
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
};
|
|
|
|
|
|
|
|
|
|
/******************************************************/
|
2004-06-24 14:18:37 +00:00
|
|
|
|
|
|
|
|
/*
|
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
|
|
|
* nm_device_test_wireless_extensions
|
2004-06-24 14:18:37 +00:00
|
|
|
*
|
|
|
|
|
* Test whether a given device is a wireless one or not.
|
|
|
|
|
*
|
|
|
|
|
*/
|
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
|
|
|
static gboolean nm_device_test_wireless_extensions (NMDevice *dev)
|
2004-06-24 14:18:37 +00:00
|
|
|
{
|
2004-10-28 19:49:55 +00:00
|
|
|
int sk;
|
2004-10-29 16:37:43 +00:00
|
|
|
int err;
|
2004-10-28 19:49:55 +00:00
|
|
|
char ioctl_buf[64];
|
2004-06-24 14:18:37 +00:00
|
|
|
|
|
|
|
|
g_return_val_if_fail (dev != NULL, FALSE);
|
2004-08-31 Dan Williams <dcbw@redhat.com>
* Remove 'debug' extern global from all files since we now
use syslog()
* src/NetworkManager.[ch]
- Break out routine that get the net.interface property from HAL,
removing that logic from nm_create_device_and_add_to_list()
- (nm_create_device_and_add_to_list): make this a bit more general so
it doesn't do the talking to HAL. Also add arguments to facilitate
the create of test devices.
- (nm_data_mark_state_changed): rename from nm_data_set_state_modified()
- (nm_data_new, main, nm_print_usage): add new argument "--enable-test-devices"
which makes NetworkManager listen for dbus commands to create test
devices, which have no backing hardware. Use when you're on a plane
for example, and/or forgot your wireless card at home. Test devices
_cannot_ be created unless NM is started with --enable-test-devices.
* src/NetworkManagerDbus.[ch]
- New "getLinkActive" method for devices
- New "setLinkActive" method for devices (only works on test devices)
- New "createTestDevice" method on NetworkManager object to create a test
device of a specified type (ie wired, wireless). UDI is created from
scratch, as is the interface name. Only works when NM is started with
--enable-test-devices switch.
- New "removeTestDevice" method on NetworkManager object which removes a
test device. Only works when NM is started with --enable-test-devices
* src/NetworkManagerDevice.[ch]
- Logic to facilitate test devices. Add variables to NMDevice struct to indicate
whether a device is a test device or not, and what its link status is.
- Deal with test devices in most functions. For those that work directly on hardware
special-case test devices.
- (nm_device_new): don't create a test device if test devices weren't enabled on the
command-line.
- (nm_device_update_link_active): split out logic for wired and wireless device link
checking to separate functions to facilitate test device link checking.
- (nm_device_set_enc_key): Since some drivers for wireless cards are daft and
don't make a distinction between System Authentication and Encryption
(namely Cisco aironet), we use Open System auth when setting a WEP key
on the card. We don't deal with Shared Key auth yet.
- (nm_device_activation_worker): split the activation cancel check logic out into
a separate routine nm_device_activation_cancel_if_needed()
- (nm_device_activation_signal_cancel): rename from nm_device_activation_cancel()
- (nm_device_fake_ap_list): Test wireless devices obviously cannot scan, so create
a list of fake access points that they can "see"
- (nm_device_is_test_device): return whether or not a device is a test device
* src/NetworkManagerPolicy.c
- (nm_policy_get_best_device): attempt to deal with wireless network selection,
previously if you "locked"/forced NM to use a wireless device but then
selected a wireless network for NM to use, it would switch to a wired device.
So, if the active device is wireless and it has a "forced" best AP, use it
if the "forced" best AP is still valid
- (nm_state_modification_monitor): deal with NULL best devices, for example
there were no usable network devices, or the last one was removed
* src/backends/NetworkManager*.c
- Deal with test devices, mostly just return success for operations like getting
a DHCP address
* test/nmtestdevices.c
- Test tool to create/remove/link-switch test devices
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@112 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-31 16:09:15 +00:00
|
|
|
|
|
|
|
|
/* We obviously cannot probe test devices (since they don't
|
|
|
|
|
* actually exist in hardware).
|
|
|
|
|
*/
|
|
|
|
|
if (dev->test_device)
|
|
|
|
|
return (FALSE);
|
|
|
|
|
|
2004-10-28 19:49:55 +00:00
|
|
|
ioctl_buf[63] = 0;
|
|
|
|
|
strncpy(ioctl_buf, nm_device_get_iface(dev), 63);
|
|
|
|
|
|
|
|
|
|
sk = iw_sockets_open ();
|
2004-10-29 16:37:43 +00:00
|
|
|
err = ioctl(sk, SIOCGIWNAME, ioctl_buf);
|
2004-10-28 19:49:55 +00:00
|
|
|
close (sk);
|
2004-10-29 16:37:43 +00:00
|
|
|
return (err == 0);
|
2004-06-24 14:18:37 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* nm_device_supports_wireless_scan
|
|
|
|
|
*
|
|
|
|
|
* Test whether a given device is a wireless one or not.
|
|
|
|
|
*
|
|
|
|
|
*/
|
|
|
|
|
static gboolean nm_device_supports_wireless_scan (NMDevice *dev)
|
|
|
|
|
{
|
2004-10-29 16:37:43 +00:00
|
|
|
int sk;
|
|
|
|
|
int err;
|
|
|
|
|
gboolean can_scan = TRUE;
|
|
|
|
|
wireless_scan_head scan_data;
|
2004-06-24 14:18:37 +00:00
|
|
|
|
|
|
|
|
g_return_val_if_fail (dev != NULL, FALSE);
|
2004-08-31 Dan Williams <dcbw@redhat.com>
* Remove 'debug' extern global from all files since we now
use syslog()
* src/NetworkManager.[ch]
- Break out routine that get the net.interface property from HAL,
removing that logic from nm_create_device_and_add_to_list()
- (nm_create_device_and_add_to_list): make this a bit more general so
it doesn't do the talking to HAL. Also add arguments to facilitate
the create of test devices.
- (nm_data_mark_state_changed): rename from nm_data_set_state_modified()
- (nm_data_new, main, nm_print_usage): add new argument "--enable-test-devices"
which makes NetworkManager listen for dbus commands to create test
devices, which have no backing hardware. Use when you're on a plane
for example, and/or forgot your wireless card at home. Test devices
_cannot_ be created unless NM is started with --enable-test-devices.
* src/NetworkManagerDbus.[ch]
- New "getLinkActive" method for devices
- New "setLinkActive" method for devices (only works on test devices)
- New "createTestDevice" method on NetworkManager object to create a test
device of a specified type (ie wired, wireless). UDI is created from
scratch, as is the interface name. Only works when NM is started with
--enable-test-devices switch.
- New "removeTestDevice" method on NetworkManager object which removes a
test device. Only works when NM is started with --enable-test-devices
* src/NetworkManagerDevice.[ch]
- Logic to facilitate test devices. Add variables to NMDevice struct to indicate
whether a device is a test device or not, and what its link status is.
- Deal with test devices in most functions. For those that work directly on hardware
special-case test devices.
- (nm_device_new): don't create a test device if test devices weren't enabled on the
command-line.
- (nm_device_update_link_active): split out logic for wired and wireless device link
checking to separate functions to facilitate test device link checking.
- (nm_device_set_enc_key): Since some drivers for wireless cards are daft and
don't make a distinction between System Authentication and Encryption
(namely Cisco aironet), we use Open System auth when setting a WEP key
on the card. We don't deal with Shared Key auth yet.
- (nm_device_activation_worker): split the activation cancel check logic out into
a separate routine nm_device_activation_cancel_if_needed()
- (nm_device_activation_signal_cancel): rename from nm_device_activation_cancel()
- (nm_device_fake_ap_list): Test wireless devices obviously cannot scan, so create
a list of fake access points that they can "see"
- (nm_device_is_test_device): return whether or not a device is a test device
* src/NetworkManagerPolicy.c
- (nm_policy_get_best_device): attempt to deal with wireless network selection,
previously if you "locked"/forced NM to use a wireless device but then
selected a wireless network for NM to use, it would switch to a wired device.
So, if the active device is wireless and it has a "forced" best AP, use it
if the "forced" best AP is still valid
- (nm_state_modification_monitor): deal with NULL best devices, for example
there were no usable network devices, or the last one was removed
* src/backends/NetworkManager*.c
- Deal with test devices, mostly just return success for operations like getting
a DHCP address
* test/nmtestdevices.c
- Test tool to create/remove/link-switch test devices
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@112 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-31 16:09:15 +00:00
|
|
|
g_return_val_if_fail (dev->type == DEVICE_TYPE_WIRELESS_ETHERNET, FALSE);
|
|
|
|
|
|
|
|
|
|
/* A test wireless device can always scan (we generate fake scan data for it) */
|
|
|
|
|
if (dev->test_device)
|
|
|
|
|
return (TRUE);
|
2004-06-24 14:18:37 +00:00
|
|
|
|
2004-10-28 19:49:55 +00:00
|
|
|
sk = iw_sockets_open ();
|
2004-10-29 16:37:43 +00:00
|
|
|
err = iw_scan (sk, (char *)nm_device_get_iface (dev), WIRELESS_EXT, &scan_data);
|
2004-06-24 14:18:37 +00:00
|
|
|
nm_dispose_scan_results (scan_data.result);
|
2004-10-29 16:37:43 +00:00
|
|
|
if ((err == -1) && (errno == EOPNOTSUPP))
|
2004-06-24 14:18:37 +00:00
|
|
|
can_scan = FALSE;
|
2004-10-28 19:49:55 +00:00
|
|
|
close (sk);
|
2004-06-24 14:18:37 +00:00
|
|
|
return (can_scan);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* nm_get_device_by_udi
|
|
|
|
|
*
|
|
|
|
|
* Search through the device list for a device with a given UDI.
|
|
|
|
|
*
|
|
|
|
|
* NOTE: the caller MUST hold the device list mutex already to make
|
|
|
|
|
* this routine thread-safe.
|
|
|
|
|
*
|
|
|
|
|
*/
|
|
|
|
|
NMDevice *nm_get_device_by_udi (NMData *data, const char *udi)
|
|
|
|
|
{
|
|
|
|
|
NMDevice *dev = NULL;
|
|
|
|
|
GSList *element;
|
|
|
|
|
|
|
|
|
|
g_return_val_if_fail (data != NULL, NULL);
|
|
|
|
|
g_return_val_if_fail (udi != NULL, NULL);
|
|
|
|
|
|
|
|
|
|
element = data->dev_list;
|
|
|
|
|
while (element)
|
|
|
|
|
{
|
|
|
|
|
dev = (NMDevice *)(element->data);
|
|
|
|
|
if (dev)
|
|
|
|
|
{
|
|
|
|
|
if (nm_null_safe_strcmp (nm_device_get_udi (dev), udi) == 0)
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
element = g_slist_next (element);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
return (dev);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* nm_get_device_by_iface
|
|
|
|
|
*
|
|
|
|
|
* Search through the device list for a device with a given iface.
|
|
|
|
|
*
|
|
|
|
|
* NOTE: the caller MUST hold the device list mutex already to make
|
|
|
|
|
* this routine thread-safe.
|
|
|
|
|
*
|
|
|
|
|
*/
|
|
|
|
|
NMDevice *nm_get_device_by_iface (NMData *data, const char *iface)
|
|
|
|
|
{
|
|
|
|
|
NMDevice *iter_dev = NULL;
|
|
|
|
|
NMDevice *found_dev = NULL;
|
|
|
|
|
GSList *element;
|
|
|
|
|
|
|
|
|
|
g_return_val_if_fail (data != NULL, NULL);
|
|
|
|
|
g_return_val_if_fail (iface != NULL, NULL);
|
|
|
|
|
|
|
|
|
|
element = data->dev_list;
|
|
|
|
|
while (element)
|
|
|
|
|
{
|
|
|
|
|
iter_dev = (NMDevice *)(element->data);
|
|
|
|
|
if (iter_dev)
|
|
|
|
|
{
|
|
|
|
|
if (nm_null_safe_strcmp (nm_device_get_iface (iter_dev), iface) == 0)
|
|
|
|
|
{
|
|
|
|
|
found_dev = iter_dev;
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
element = g_slist_next (element);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
return (found_dev);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
/*****************************************************************************/
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
/* NMDevice object routines */
|
2004-06-24 14:18:37 +00:00
|
|
|
/*****************************************************************************/
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* nm_device_new
|
|
|
|
|
*
|
2004-08-31 Dan Williams <dcbw@redhat.com>
* Remove 'debug' extern global from all files since we now
use syslog()
* src/NetworkManager.[ch]
- Break out routine that get the net.interface property from HAL,
removing that logic from nm_create_device_and_add_to_list()
- (nm_create_device_and_add_to_list): make this a bit more general so
it doesn't do the talking to HAL. Also add arguments to facilitate
the create of test devices.
- (nm_data_mark_state_changed): rename from nm_data_set_state_modified()
- (nm_data_new, main, nm_print_usage): add new argument "--enable-test-devices"
which makes NetworkManager listen for dbus commands to create test
devices, which have no backing hardware. Use when you're on a plane
for example, and/or forgot your wireless card at home. Test devices
_cannot_ be created unless NM is started with --enable-test-devices.
* src/NetworkManagerDbus.[ch]
- New "getLinkActive" method for devices
- New "setLinkActive" method for devices (only works on test devices)
- New "createTestDevice" method on NetworkManager object to create a test
device of a specified type (ie wired, wireless). UDI is created from
scratch, as is the interface name. Only works when NM is started with
--enable-test-devices switch.
- New "removeTestDevice" method on NetworkManager object which removes a
test device. Only works when NM is started with --enable-test-devices
* src/NetworkManagerDevice.[ch]
- Logic to facilitate test devices. Add variables to NMDevice struct to indicate
whether a device is a test device or not, and what its link status is.
- Deal with test devices in most functions. For those that work directly on hardware
special-case test devices.
- (nm_device_new): don't create a test device if test devices weren't enabled on the
command-line.
- (nm_device_update_link_active): split out logic for wired and wireless device link
checking to separate functions to facilitate test device link checking.
- (nm_device_set_enc_key): Since some drivers for wireless cards are daft and
don't make a distinction between System Authentication and Encryption
(namely Cisco aironet), we use Open System auth when setting a WEP key
on the card. We don't deal with Shared Key auth yet.
- (nm_device_activation_worker): split the activation cancel check logic out into
a separate routine nm_device_activation_cancel_if_needed()
- (nm_device_activation_signal_cancel): rename from nm_device_activation_cancel()
- (nm_device_fake_ap_list): Test wireless devices obviously cannot scan, so create
a list of fake access points that they can "see"
- (nm_device_is_test_device): return whether or not a device is a test device
* src/NetworkManagerPolicy.c
- (nm_policy_get_best_device): attempt to deal with wireless network selection,
previously if you "locked"/forced NM to use a wireless device but then
selected a wireless network for NM to use, it would switch to a wired device.
So, if the active device is wireless and it has a "forced" best AP, use it
if the "forced" best AP is still valid
- (nm_state_modification_monitor): deal with NULL best devices, for example
there were no usable network devices, or the last one was removed
* src/backends/NetworkManager*.c
- Deal with test devices, mostly just return success for operations like getting
a DHCP address
* test/nmtestdevices.c
- Test tool to create/remove/link-switch test devices
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@112 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-31 16:09:15 +00:00
|
|
|
* Creates and initializes the structure representation of an NM device. For test
|
|
|
|
|
* devices, a device type other than DEVICE_TYPE_DONT_KNOW must be specified, this
|
|
|
|
|
* argument is ignored for real hardware devices since they are auto-probed.
|
2004-06-24 14:18:37 +00:00
|
|
|
*
|
|
|
|
|
*/
|
2004-10-21 17:42:14 +00:00
|
|
|
NMDevice *nm_device_new (const char *iface, const char *udi, gboolean test_dev, NMDeviceType test_dev_type, NMData *app_data)
|
2004-06-24 14:18:37 +00:00
|
|
|
{
|
|
|
|
|
NMDevice *dev;
|
|
|
|
|
|
|
|
|
|
g_return_val_if_fail (iface != NULL, NULL);
|
2004-08-24 00:31:47 +00:00
|
|
|
g_return_val_if_fail (strlen (iface) > 0, NULL);
|
2004-10-21 17:42:14 +00:00
|
|
|
g_return_val_if_fail (app_data != NULL, NULL);
|
2004-08-24 00:31:47 +00:00
|
|
|
|
2004-08-31 Dan Williams <dcbw@redhat.com>
* Remove 'debug' extern global from all files since we now
use syslog()
* src/NetworkManager.[ch]
- Break out routine that get the net.interface property from HAL,
removing that logic from nm_create_device_and_add_to_list()
- (nm_create_device_and_add_to_list): make this a bit more general so
it doesn't do the talking to HAL. Also add arguments to facilitate
the create of test devices.
- (nm_data_mark_state_changed): rename from nm_data_set_state_modified()
- (nm_data_new, main, nm_print_usage): add new argument "--enable-test-devices"
which makes NetworkManager listen for dbus commands to create test
devices, which have no backing hardware. Use when you're on a plane
for example, and/or forgot your wireless card at home. Test devices
_cannot_ be created unless NM is started with --enable-test-devices.
* src/NetworkManagerDbus.[ch]
- New "getLinkActive" method for devices
- New "setLinkActive" method for devices (only works on test devices)
- New "createTestDevice" method on NetworkManager object to create a test
device of a specified type (ie wired, wireless). UDI is created from
scratch, as is the interface name. Only works when NM is started with
--enable-test-devices switch.
- New "removeTestDevice" method on NetworkManager object which removes a
test device. Only works when NM is started with --enable-test-devices
* src/NetworkManagerDevice.[ch]
- Logic to facilitate test devices. Add variables to NMDevice struct to indicate
whether a device is a test device or not, and what its link status is.
- Deal with test devices in most functions. For those that work directly on hardware
special-case test devices.
- (nm_device_new): don't create a test device if test devices weren't enabled on the
command-line.
- (nm_device_update_link_active): split out logic for wired and wireless device link
checking to separate functions to facilitate test device link checking.
- (nm_device_set_enc_key): Since some drivers for wireless cards are daft and
don't make a distinction between System Authentication and Encryption
(namely Cisco aironet), we use Open System auth when setting a WEP key
on the card. We don't deal with Shared Key auth yet.
- (nm_device_activation_worker): split the activation cancel check logic out into
a separate routine nm_device_activation_cancel_if_needed()
- (nm_device_activation_signal_cancel): rename from nm_device_activation_cancel()
- (nm_device_fake_ap_list): Test wireless devices obviously cannot scan, so create
a list of fake access points that they can "see"
- (nm_device_is_test_device): return whether or not a device is a test device
* src/NetworkManagerPolicy.c
- (nm_policy_get_best_device): attempt to deal with wireless network selection,
previously if you "locked"/forced NM to use a wireless device but then
selected a wireless network for NM to use, it would switch to a wired device.
So, if the active device is wireless and it has a "forced" best AP, use it
if the "forced" best AP is still valid
- (nm_state_modification_monitor): deal with NULL best devices, for example
there were no usable network devices, or the last one was removed
* src/backends/NetworkManager*.c
- Deal with test devices, mostly just return success for operations like getting
a DHCP address
* test/nmtestdevices.c
- Test tool to create/remove/link-switch test devices
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@112 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-31 16:09:15 +00:00
|
|
|
/* Test devices must have a valid type specified */
|
|
|
|
|
if (test_dev && !(test_dev_type != DEVICE_TYPE_DONT_KNOW))
|
|
|
|
|
return (NULL);
|
|
|
|
|
|
|
|
|
|
/* Another check to make sure we don't create a test device unless
|
|
|
|
|
* test devices were enabled on the command line.
|
|
|
|
|
*/
|
2004-10-21 17:42:14 +00:00
|
|
|
if (!app_data->enable_test_devices && test_dev)
|
2004-08-31 Dan Williams <dcbw@redhat.com>
* Remove 'debug' extern global from all files since we now
use syslog()
* src/NetworkManager.[ch]
- Break out routine that get the net.interface property from HAL,
removing that logic from nm_create_device_and_add_to_list()
- (nm_create_device_and_add_to_list): make this a bit more general so
it doesn't do the talking to HAL. Also add arguments to facilitate
the create of test devices.
- (nm_data_mark_state_changed): rename from nm_data_set_state_modified()
- (nm_data_new, main, nm_print_usage): add new argument "--enable-test-devices"
which makes NetworkManager listen for dbus commands to create test
devices, which have no backing hardware. Use when you're on a plane
for example, and/or forgot your wireless card at home. Test devices
_cannot_ be created unless NM is started with --enable-test-devices.
* src/NetworkManagerDbus.[ch]
- New "getLinkActive" method for devices
- New "setLinkActive" method for devices (only works on test devices)
- New "createTestDevice" method on NetworkManager object to create a test
device of a specified type (ie wired, wireless). UDI is created from
scratch, as is the interface name. Only works when NM is started with
--enable-test-devices switch.
- New "removeTestDevice" method on NetworkManager object which removes a
test device. Only works when NM is started with --enable-test-devices
* src/NetworkManagerDevice.[ch]
- Logic to facilitate test devices. Add variables to NMDevice struct to indicate
whether a device is a test device or not, and what its link status is.
- Deal with test devices in most functions. For those that work directly on hardware
special-case test devices.
- (nm_device_new): don't create a test device if test devices weren't enabled on the
command-line.
- (nm_device_update_link_active): split out logic for wired and wireless device link
checking to separate functions to facilitate test device link checking.
- (nm_device_set_enc_key): Since some drivers for wireless cards are daft and
don't make a distinction between System Authentication and Encryption
(namely Cisco aironet), we use Open System auth when setting a WEP key
on the card. We don't deal with Shared Key auth yet.
- (nm_device_activation_worker): split the activation cancel check logic out into
a separate routine nm_device_activation_cancel_if_needed()
- (nm_device_activation_signal_cancel): rename from nm_device_activation_cancel()
- (nm_device_fake_ap_list): Test wireless devices obviously cannot scan, so create
a list of fake access points that they can "see"
- (nm_device_is_test_device): return whether or not a device is a test device
* src/NetworkManagerPolicy.c
- (nm_policy_get_best_device): attempt to deal with wireless network selection,
previously if you "locked"/forced NM to use a wireless device but then
selected a wireless network for NM to use, it would switch to a wired device.
So, if the active device is wireless and it has a "forced" best AP, use it
if the "forced" best AP is still valid
- (nm_state_modification_monitor): deal with NULL best devices, for example
there were no usable network devices, or the last one was removed
* src/backends/NetworkManager*.c
- Deal with test devices, mostly just return success for operations like getting
a DHCP address
* test/nmtestdevices.c
- Test tool to create/remove/link-switch test devices
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@112 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-31 16:09:15 +00:00
|
|
|
{
|
|
|
|
|
syslog (LOG_ERR, "nm_device_new(): attempt to create a test device, but test devices were not enabled"
|
|
|
|
|
" on the command line. Will not create the device.\n");
|
|
|
|
|
return (NULL);
|
|
|
|
|
}
|
|
|
|
|
|
2004-06-24 14:18:37 +00:00
|
|
|
dev = g_new0 (NMDevice, 1);
|
|
|
|
|
if (!dev)
|
|
|
|
|
{
|
2004-08-23 19:20:49 +00:00
|
|
|
syslog (LOG_ERR, "nm_device_new() could not allocate a new device... Not enough memory?");
|
2004-06-24 14:18:37 +00:00
|
|
|
return (NULL);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
dev->refcount = 1;
|
2004-07-25 02:40:19 +00:00
|
|
|
dev->app_data = app_data;
|
2004-06-24 14:18:37 +00:00
|
|
|
dev->iface = g_strdup (iface);
|
2004-08-31 Dan Williams <dcbw@redhat.com>
* Remove 'debug' extern global from all files since we now
use syslog()
* src/NetworkManager.[ch]
- Break out routine that get the net.interface property from HAL,
removing that logic from nm_create_device_and_add_to_list()
- (nm_create_device_and_add_to_list): make this a bit more general so
it doesn't do the talking to HAL. Also add arguments to facilitate
the create of test devices.
- (nm_data_mark_state_changed): rename from nm_data_set_state_modified()
- (nm_data_new, main, nm_print_usage): add new argument "--enable-test-devices"
which makes NetworkManager listen for dbus commands to create test
devices, which have no backing hardware. Use when you're on a plane
for example, and/or forgot your wireless card at home. Test devices
_cannot_ be created unless NM is started with --enable-test-devices.
* src/NetworkManagerDbus.[ch]
- New "getLinkActive" method for devices
- New "setLinkActive" method for devices (only works on test devices)
- New "createTestDevice" method on NetworkManager object to create a test
device of a specified type (ie wired, wireless). UDI is created from
scratch, as is the interface name. Only works when NM is started with
--enable-test-devices switch.
- New "removeTestDevice" method on NetworkManager object which removes a
test device. Only works when NM is started with --enable-test-devices
* src/NetworkManagerDevice.[ch]
- Logic to facilitate test devices. Add variables to NMDevice struct to indicate
whether a device is a test device or not, and what its link status is.
- Deal with test devices in most functions. For those that work directly on hardware
special-case test devices.
- (nm_device_new): don't create a test device if test devices weren't enabled on the
command-line.
- (nm_device_update_link_active): split out logic for wired and wireless device link
checking to separate functions to facilitate test device link checking.
- (nm_device_set_enc_key): Since some drivers for wireless cards are daft and
don't make a distinction between System Authentication and Encryption
(namely Cisco aironet), we use Open System auth when setting a WEP key
on the card. We don't deal with Shared Key auth yet.
- (nm_device_activation_worker): split the activation cancel check logic out into
a separate routine nm_device_activation_cancel_if_needed()
- (nm_device_activation_signal_cancel): rename from nm_device_activation_cancel()
- (nm_device_fake_ap_list): Test wireless devices obviously cannot scan, so create
a list of fake access points that they can "see"
- (nm_device_is_test_device): return whether or not a device is a test device
* src/NetworkManagerPolicy.c
- (nm_policy_get_best_device): attempt to deal with wireless network selection,
previously if you "locked"/forced NM to use a wireless device but then
selected a wireless network for NM to use, it would switch to a wired device.
So, if the active device is wireless and it has a "forced" best AP, use it
if the "forced" best AP is still valid
- (nm_state_modification_monitor): deal with NULL best devices, for example
there were no usable network devices, or the last one was removed
* src/backends/NetworkManager*.c
- Deal with test devices, mostly just return success for operations like getting
a DHCP address
* test/nmtestdevices.c
- Test tool to create/remove/link-switch test devices
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@112 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-31 16:09:15 +00:00
|
|
|
dev->test_device = test_dev;
|
2004-10-21 17:42:14 +00:00
|
|
|
nm_device_set_udi (dev, udi);
|
2004-08-31 Dan Williams <dcbw@redhat.com>
* Remove 'debug' extern global from all files since we now
use syslog()
* src/NetworkManager.[ch]
- Break out routine that get the net.interface property from HAL,
removing that logic from nm_create_device_and_add_to_list()
- (nm_create_device_and_add_to_list): make this a bit more general so
it doesn't do the talking to HAL. Also add arguments to facilitate
the create of test devices.
- (nm_data_mark_state_changed): rename from nm_data_set_state_modified()
- (nm_data_new, main, nm_print_usage): add new argument "--enable-test-devices"
which makes NetworkManager listen for dbus commands to create test
devices, which have no backing hardware. Use when you're on a plane
for example, and/or forgot your wireless card at home. Test devices
_cannot_ be created unless NM is started with --enable-test-devices.
* src/NetworkManagerDbus.[ch]
- New "getLinkActive" method for devices
- New "setLinkActive" method for devices (only works on test devices)
- New "createTestDevice" method on NetworkManager object to create a test
device of a specified type (ie wired, wireless). UDI is created from
scratch, as is the interface name. Only works when NM is started with
--enable-test-devices switch.
- New "removeTestDevice" method on NetworkManager object which removes a
test device. Only works when NM is started with --enable-test-devices
* src/NetworkManagerDevice.[ch]
- Logic to facilitate test devices. Add variables to NMDevice struct to indicate
whether a device is a test device or not, and what its link status is.
- Deal with test devices in most functions. For those that work directly on hardware
special-case test devices.
- (nm_device_new): don't create a test device if test devices weren't enabled on the
command-line.
- (nm_device_update_link_active): split out logic for wired and wireless device link
checking to separate functions to facilitate test device link checking.
- (nm_device_set_enc_key): Since some drivers for wireless cards are daft and
don't make a distinction between System Authentication and Encryption
(namely Cisco aironet), we use Open System auth when setting a WEP key
on the card. We don't deal with Shared Key auth yet.
- (nm_device_activation_worker): split the activation cancel check logic out into
a separate routine nm_device_activation_cancel_if_needed()
- (nm_device_activation_signal_cancel): rename from nm_device_activation_cancel()
- (nm_device_fake_ap_list): Test wireless devices obviously cannot scan, so create
a list of fake access points that they can "see"
- (nm_device_is_test_device): return whether or not a device is a test device
* src/NetworkManagerPolicy.c
- (nm_policy_get_best_device): attempt to deal with wireless network selection,
previously if you "locked"/forced NM to use a wireless device but then
selected a wireless network for NM to use, it would switch to a wired device.
So, if the active device is wireless and it has a "forced" best AP, use it
if the "forced" best AP is still valid
- (nm_state_modification_monitor): deal with NULL best devices, for example
there were no usable network devices, or the last one was removed
* src/backends/NetworkManager*.c
- Deal with test devices, mostly just return success for operations like getting
a DHCP address
* test/nmtestdevices.c
- Test tool to create/remove/link-switch test devices
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@112 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-31 16:09:15 +00:00
|
|
|
|
|
|
|
|
/* Real hardware devices are probed for their type, test devices must have
|
|
|
|
|
* their type specified.
|
|
|
|
|
*/
|
|
|
|
|
if (test_dev)
|
|
|
|
|
dev->type = test_dev_type;
|
|
|
|
|
else
|
|
|
|
|
dev->type = nm_device_test_wireless_extensions (dev) ?
|
2004-08-05 18:54:29 +00:00
|
|
|
DEVICE_TYPE_WIRELESS_ETHERNET : DEVICE_TYPE_WIRED_ETHERNET;
|
2004-06-24 14:18:37 +00:00
|
|
|
|
2004-09-08 18:14:42 +00:00
|
|
|
/* Initialize wireless-specific options */
|
2004-08-05 18:54:29 +00:00
|
|
|
if (nm_device_is_wireless (dev))
|
2004-06-24 14:18:37 +00:00
|
|
|
{
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
if (!(dev->options.wireless.scan_mutex = g_mutex_new ()))
|
|
|
|
|
{
|
|
|
|
|
g_free (dev->iface);
|
|
|
|
|
g_free (dev);
|
|
|
|
|
return (NULL);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (!(dev->options.wireless.best_ap_mutex = g_mutex_new ()))
|
|
|
|
|
{
|
|
|
|
|
g_mutex_free (dev->options.wireless.scan_mutex);
|
|
|
|
|
g_free (dev->iface);
|
|
|
|
|
g_free (dev);
|
|
|
|
|
return (NULL);
|
|
|
|
|
}
|
|
|
|
|
|
2004-08-05 18:54:29 +00:00
|
|
|
if (!(dev->options.wireless.ap_list = nm_ap_list_new (NETWORK_TYPE_DEVICE)))
|
2004-06-24 14:18:37 +00:00
|
|
|
{
|
|
|
|
|
g_free (dev->iface);
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
g_mutex_free (dev->options.wireless.best_ap_mutex);
|
2004-08-05 18:54:29 +00:00
|
|
|
g_free (dev);
|
2004-06-24 14:18:37 +00:00
|
|
|
return (NULL);
|
|
|
|
|
}
|
2004-08-05 18:54:29 +00:00
|
|
|
dev->options.wireless.supports_wireless_scan = nm_device_supports_wireless_scan (dev);
|
2004-09-12 03:06:44 +00:00
|
|
|
|
|
|
|
|
/* Perform an initial wireless scan */
|
2004-10-12 16:44:58 +00:00
|
|
|
nm_device_set_mode_managed (dev);
|
2004-09-12 03:06:44 +00:00
|
|
|
nm_device_do_wireless_scan (dev);
|
|
|
|
|
nm_device_update_best_ap (dev);
|
2004-06-24 14:18:37 +00:00
|
|
|
}
|
|
|
|
|
|
2004-10-21 17:42:14 +00:00
|
|
|
dev->driver_support_level = nm_get_driver_support_level (dev->app_data->hal_ctx, dev);
|
|
|
|
|
|
2004-10-21 19:00:13 +00:00
|
|
|
if (nm_device_get_driver_support_level (dev) != NM_DRIVER_UNSUPPORTED)
|
|
|
|
|
{
|
|
|
|
|
/* Grab IP config data for this device from the system configuration files */
|
|
|
|
|
nm_device_update_ip4_address (dev);
|
|
|
|
|
nm_system_device_update_config_info (dev);
|
2004-09-08 18:14:42 +00:00
|
|
|
|
2004-10-21 19:00:13 +00:00
|
|
|
/* Have to bring the device up before checking link status. */
|
|
|
|
|
nm_device_bring_up (dev);
|
|
|
|
|
nm_device_update_link_active (dev, TRUE);
|
|
|
|
|
}
|
2004-06-24 14:18:37 +00:00
|
|
|
|
|
|
|
|
return (dev);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* Refcounting functions
|
|
|
|
|
*/
|
|
|
|
|
void nm_device_ref (NMDevice *dev)
|
|
|
|
|
{
|
|
|
|
|
g_return_if_fail (dev != NULL);
|
|
|
|
|
|
|
|
|
|
dev->refcount++;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
void nm_device_unref (NMDevice *dev)
|
|
|
|
|
{
|
|
|
|
|
g_return_if_fail (dev != NULL);
|
|
|
|
|
|
|
|
|
|
dev->refcount--;
|
2004-08-05 18:54:29 +00:00
|
|
|
if (dev->refcount <= 0)
|
2004-06-24 14:18:37 +00:00
|
|
|
{
|
|
|
|
|
nm_device_ap_list_clear (dev);
|
2004-08-05 18:54:29 +00:00
|
|
|
dev->options.wireless.ap_list = NULL;
|
2004-06-24 14:18:37 +00:00
|
|
|
|
|
|
|
|
g_free (dev->udi);
|
|
|
|
|
g_free (dev->iface);
|
2004-08-05 18:54:29 +00:00
|
|
|
if (nm_device_is_wireless (dev))
|
2004-06-24 14:18:37 +00:00
|
|
|
{
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
g_mutex_free (dev->options.wireless.scan_mutex);
|
2004-10-11 21:32:19 +00:00
|
|
|
if (dev->options.wireless.ap_list)
|
|
|
|
|
nm_ap_list_unref (dev->options.wireless.ap_list);
|
|
|
|
|
if (dev->options.wireless.cached_ap_list1)
|
|
|
|
|
nm_ap_list_unref (dev->options.wireless.cached_ap_list1);
|
|
|
|
|
if (dev->options.wireless.cached_ap_list2)
|
|
|
|
|
nm_ap_list_unref (dev->options.wireless.cached_ap_list2);
|
|
|
|
|
if (dev->options.wireless.cached_ap_list3)
|
|
|
|
|
nm_ap_list_unref (dev->options.wireless.cached_ap_list3);
|
2004-08-05 18:54:29 +00:00
|
|
|
nm_ap_unref (dev->options.wireless.best_ap);
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
g_mutex_free (dev->options.wireless.best_ap_mutex);
|
2004-06-24 14:18:37 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
dev->udi = NULL;
|
|
|
|
|
dev->iface = NULL;
|
2004-08-05 18:54:29 +00:00
|
|
|
g_free (dev);
|
2004-06-24 14:18:37 +00:00
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* Get/set functions for UDI
|
|
|
|
|
*/
|
|
|
|
|
char * nm_device_get_udi (NMDevice *dev)
|
|
|
|
|
{
|
|
|
|
|
g_return_val_if_fail (dev != NULL, NULL);
|
|
|
|
|
|
|
|
|
|
return (dev->udi);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
void nm_device_set_udi (NMDevice *dev, const char *udi)
|
|
|
|
|
{
|
|
|
|
|
g_return_if_fail (dev != NULL);
|
|
|
|
|
g_return_if_fail (udi != NULL);
|
|
|
|
|
|
|
|
|
|
if (dev->udi)
|
|
|
|
|
g_free (dev->udi);
|
|
|
|
|
|
|
|
|
|
dev->udi = g_strdup (udi);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* Get/set functions for iface
|
|
|
|
|
*/
|
2004-10-21 17:42:14 +00:00
|
|
|
const char * nm_device_get_iface (NMDevice *dev)
|
2004-06-24 14:18:37 +00:00
|
|
|
{
|
|
|
|
|
g_return_val_if_fail (dev != NULL, NULL);
|
|
|
|
|
|
|
|
|
|
return (dev->iface);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
2004-08-05 18:54:29 +00:00
|
|
|
* Get/set functions for type
|
2004-06-24 14:18:37 +00:00
|
|
|
*/
|
2004-08-05 18:54:29 +00:00
|
|
|
guint nm_device_get_type (NMDevice *dev)
|
2004-06-24 14:18:37 +00:00
|
|
|
{
|
2004-08-05 18:54:29 +00:00
|
|
|
g_return_val_if_fail (dev != NULL, DEVICE_TYPE_DONT_KNOW);
|
2004-06-24 14:18:37 +00:00
|
|
|
|
2004-08-05 18:54:29 +00:00
|
|
|
return (dev->type);
|
2004-06-24 14:18:37 +00:00
|
|
|
}
|
|
|
|
|
|
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
|
|
|
gboolean nm_device_is_wireless (NMDevice *dev)
|
|
|
|
|
{
|
|
|
|
|
g_return_val_if_fail (dev != NULL, FALSE);
|
|
|
|
|
|
2004-08-05 18:54:29 +00:00
|
|
|
return (dev->type == DEVICE_TYPE_WIRELESS_ETHERNET);
|
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
|
|
|
}
|
|
|
|
|
|
|
|
|
|
gboolean nm_device_is_wired (NMDevice *dev)
|
|
|
|
|
{
|
|
|
|
|
g_return_val_if_fail (dev != NULL, FALSE);
|
|
|
|
|
|
2004-08-05 18:54:29 +00:00
|
|
|
return (dev->type == DEVICE_TYPE_WIRED_ETHERNET);
|
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
|
|
|
}
|
|
|
|
|
|
2004-06-24 14:18:37 +00:00
|
|
|
|
2004-10-21 17:42:14 +00:00
|
|
|
/*
|
|
|
|
|
* Accessor for driver support level
|
|
|
|
|
*/
|
|
|
|
|
NMDriverSupportLevel nm_device_get_driver_support_level (NMDevice *dev)
|
|
|
|
|
{
|
|
|
|
|
g_return_val_if_fail (dev != NULL, NM_DRIVER_UNSUPPORTED);
|
|
|
|
|
|
|
|
|
|
return (dev->driver_support_level);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
2004-06-24 14:18:37 +00:00
|
|
|
/*
|
|
|
|
|
* Get/set functions for link_active
|
|
|
|
|
*/
|
|
|
|
|
gboolean nm_device_get_link_active (NMDevice *dev)
|
|
|
|
|
{
|
|
|
|
|
g_return_val_if_fail (dev != NULL, FALSE);
|
|
|
|
|
|
|
|
|
|
return (dev->link_active);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
void nm_device_set_link_active (NMDevice *dev, const gboolean link_active)
|
|
|
|
|
{
|
|
|
|
|
g_return_if_fail (dev != NULL);
|
|
|
|
|
|
|
|
|
|
dev->link_active = link_active;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* Get function for supports_wireless_scan
|
|
|
|
|
*/
|
|
|
|
|
gboolean nm_device_get_supports_wireless_scan (NMDevice *dev)
|
|
|
|
|
{
|
|
|
|
|
g_return_val_if_fail (dev != NULL, FALSE);
|
|
|
|
|
|
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
|
|
|
if (!nm_device_is_wireless (dev))
|
|
|
|
|
return (FALSE);
|
|
|
|
|
|
2004-08-05 18:54:29 +00:00
|
|
|
return (dev->options.wireless.supports_wireless_scan);
|
2004-06-24 14:18:37 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
2004-08-31 Dan Williams <dcbw@redhat.com>
* Remove 'debug' extern global from all files since we now
use syslog()
* src/NetworkManager.[ch]
- Break out routine that get the net.interface property from HAL,
removing that logic from nm_create_device_and_add_to_list()
- (nm_create_device_and_add_to_list): make this a bit more general so
it doesn't do the talking to HAL. Also add arguments to facilitate
the create of test devices.
- (nm_data_mark_state_changed): rename from nm_data_set_state_modified()
- (nm_data_new, main, nm_print_usage): add new argument "--enable-test-devices"
which makes NetworkManager listen for dbus commands to create test
devices, which have no backing hardware. Use when you're on a plane
for example, and/or forgot your wireless card at home. Test devices
_cannot_ be created unless NM is started with --enable-test-devices.
* src/NetworkManagerDbus.[ch]
- New "getLinkActive" method for devices
- New "setLinkActive" method for devices (only works on test devices)
- New "createTestDevice" method on NetworkManager object to create a test
device of a specified type (ie wired, wireless). UDI is created from
scratch, as is the interface name. Only works when NM is started with
--enable-test-devices switch.
- New "removeTestDevice" method on NetworkManager object which removes a
test device. Only works when NM is started with --enable-test-devices
* src/NetworkManagerDevice.[ch]
- Logic to facilitate test devices. Add variables to NMDevice struct to indicate
whether a device is a test device or not, and what its link status is.
- Deal with test devices in most functions. For those that work directly on hardware
special-case test devices.
- (nm_device_new): don't create a test device if test devices weren't enabled on the
command-line.
- (nm_device_update_link_active): split out logic for wired and wireless device link
checking to separate functions to facilitate test device link checking.
- (nm_device_set_enc_key): Since some drivers for wireless cards are daft and
don't make a distinction between System Authentication and Encryption
(namely Cisco aironet), we use Open System auth when setting a WEP key
on the card. We don't deal with Shared Key auth yet.
- (nm_device_activation_worker): split the activation cancel check logic out into
a separate routine nm_device_activation_cancel_if_needed()
- (nm_device_activation_signal_cancel): rename from nm_device_activation_cancel()
- (nm_device_fake_ap_list): Test wireless devices obviously cannot scan, so create
a list of fake access points that they can "see"
- (nm_device_is_test_device): return whether or not a device is a test device
* src/NetworkManagerPolicy.c
- (nm_policy_get_best_device): attempt to deal with wireless network selection,
previously if you "locked"/forced NM to use a wireless device but then
selected a wireless network for NM to use, it would switch to a wired device.
So, if the active device is wireless and it has a "forced" best AP, use it
if the "forced" best AP is still valid
- (nm_state_modification_monitor): deal with NULL best devices, for example
there were no usable network devices, or the last one was removed
* src/backends/NetworkManager*.c
- Deal with test devices, mostly just return success for operations like getting
a DHCP address
* test/nmtestdevices.c
- Test tool to create/remove/link-switch test devices
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@112 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-31 16:09:15 +00:00
|
|
|
* nm_device_wireless_link_active
|
2004-06-24 14:18:37 +00:00
|
|
|
*
|
2004-08-31 Dan Williams <dcbw@redhat.com>
* Remove 'debug' extern global from all files since we now
use syslog()
* src/NetworkManager.[ch]
- Break out routine that get the net.interface property from HAL,
removing that logic from nm_create_device_and_add_to_list()
- (nm_create_device_and_add_to_list): make this a bit more general so
it doesn't do the talking to HAL. Also add arguments to facilitate
the create of test devices.
- (nm_data_mark_state_changed): rename from nm_data_set_state_modified()
- (nm_data_new, main, nm_print_usage): add new argument "--enable-test-devices"
which makes NetworkManager listen for dbus commands to create test
devices, which have no backing hardware. Use when you're on a plane
for example, and/or forgot your wireless card at home. Test devices
_cannot_ be created unless NM is started with --enable-test-devices.
* src/NetworkManagerDbus.[ch]
- New "getLinkActive" method for devices
- New "setLinkActive" method for devices (only works on test devices)
- New "createTestDevice" method on NetworkManager object to create a test
device of a specified type (ie wired, wireless). UDI is created from
scratch, as is the interface name. Only works when NM is started with
--enable-test-devices switch.
- New "removeTestDevice" method on NetworkManager object which removes a
test device. Only works when NM is started with --enable-test-devices
* src/NetworkManagerDevice.[ch]
- Logic to facilitate test devices. Add variables to NMDevice struct to indicate
whether a device is a test device or not, and what its link status is.
- Deal with test devices in most functions. For those that work directly on hardware
special-case test devices.
- (nm_device_new): don't create a test device if test devices weren't enabled on the
command-line.
- (nm_device_update_link_active): split out logic for wired and wireless device link
checking to separate functions to facilitate test device link checking.
- (nm_device_set_enc_key): Since some drivers for wireless cards are daft and
don't make a distinction between System Authentication and Encryption
(namely Cisco aironet), we use Open System auth when setting a WEP key
on the card. We don't deal with Shared Key auth yet.
- (nm_device_activation_worker): split the activation cancel check logic out into
a separate routine nm_device_activation_cancel_if_needed()
- (nm_device_activation_signal_cancel): rename from nm_device_activation_cancel()
- (nm_device_fake_ap_list): Test wireless devices obviously cannot scan, so create
a list of fake access points that they can "see"
- (nm_device_is_test_device): return whether or not a device is a test device
* src/NetworkManagerPolicy.c
- (nm_policy_get_best_device): attempt to deal with wireless network selection,
previously if you "locked"/forced NM to use a wireless device but then
selected a wireless network for NM to use, it would switch to a wired device.
So, if the active device is wireless and it has a "forced" best AP, use it
if the "forced" best AP is still valid
- (nm_state_modification_monitor): deal with NULL best devices, for example
there were no usable network devices, or the last one was removed
* src/backends/NetworkManager*.c
- Deal with test devices, mostly just return success for operations like getting
a DHCP address
* test/nmtestdevices.c
- Test tool to create/remove/link-switch test devices
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@112 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-31 16:09:15 +00:00
|
|
|
* Gets the link state of a wireless device
|
2004-06-24 14:18:37 +00:00
|
|
|
*
|
|
|
|
|
*/
|
2004-08-31 Dan Williams <dcbw@redhat.com>
* Remove 'debug' extern global from all files since we now
use syslog()
* src/NetworkManager.[ch]
- Break out routine that get the net.interface property from HAL,
removing that logic from nm_create_device_and_add_to_list()
- (nm_create_device_and_add_to_list): make this a bit more general so
it doesn't do the talking to HAL. Also add arguments to facilitate
the create of test devices.
- (nm_data_mark_state_changed): rename from nm_data_set_state_modified()
- (nm_data_new, main, nm_print_usage): add new argument "--enable-test-devices"
which makes NetworkManager listen for dbus commands to create test
devices, which have no backing hardware. Use when you're on a plane
for example, and/or forgot your wireless card at home. Test devices
_cannot_ be created unless NM is started with --enable-test-devices.
* src/NetworkManagerDbus.[ch]
- New "getLinkActive" method for devices
- New "setLinkActive" method for devices (only works on test devices)
- New "createTestDevice" method on NetworkManager object to create a test
device of a specified type (ie wired, wireless). UDI is created from
scratch, as is the interface name. Only works when NM is started with
--enable-test-devices switch.
- New "removeTestDevice" method on NetworkManager object which removes a
test device. Only works when NM is started with --enable-test-devices
* src/NetworkManagerDevice.[ch]
- Logic to facilitate test devices. Add variables to NMDevice struct to indicate
whether a device is a test device or not, and what its link status is.
- Deal with test devices in most functions. For those that work directly on hardware
special-case test devices.
- (nm_device_new): don't create a test device if test devices weren't enabled on the
command-line.
- (nm_device_update_link_active): split out logic for wired and wireless device link
checking to separate functions to facilitate test device link checking.
- (nm_device_set_enc_key): Since some drivers for wireless cards are daft and
don't make a distinction between System Authentication and Encryption
(namely Cisco aironet), we use Open System auth when setting a WEP key
on the card. We don't deal with Shared Key auth yet.
- (nm_device_activation_worker): split the activation cancel check logic out into
a separate routine nm_device_activation_cancel_if_needed()
- (nm_device_activation_signal_cancel): rename from nm_device_activation_cancel()
- (nm_device_fake_ap_list): Test wireless devices obviously cannot scan, so create
a list of fake access points that they can "see"
- (nm_device_is_test_device): return whether or not a device is a test device
* src/NetworkManagerPolicy.c
- (nm_policy_get_best_device): attempt to deal with wireless network selection,
previously if you "locked"/forced NM to use a wireless device but then
selected a wireless network for NM to use, it would switch to a wired device.
So, if the active device is wireless and it has a "forced" best AP, use it
if the "forced" best AP is still valid
- (nm_state_modification_monitor): deal with NULL best devices, for example
there were no usable network devices, or the last one was removed
* src/backends/NetworkManager*.c
- Deal with test devices, mostly just return success for operations like getting
a DHCP address
* test/nmtestdevices.c
- Test tool to create/remove/link-switch test devices
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@112 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-31 16:09:15 +00:00
|
|
|
static gboolean nm_device_wireless_link_active (NMDevice *dev)
|
2004-06-24 14:18:37 +00:00
|
|
|
{
|
2004-08-31 Dan Williams <dcbw@redhat.com>
* Remove 'debug' extern global from all files since we now
use syslog()
* src/NetworkManager.[ch]
- Break out routine that get the net.interface property from HAL,
removing that logic from nm_create_device_and_add_to_list()
- (nm_create_device_and_add_to_list): make this a bit more general so
it doesn't do the talking to HAL. Also add arguments to facilitate
the create of test devices.
- (nm_data_mark_state_changed): rename from nm_data_set_state_modified()
- (nm_data_new, main, nm_print_usage): add new argument "--enable-test-devices"
which makes NetworkManager listen for dbus commands to create test
devices, which have no backing hardware. Use when you're on a plane
for example, and/or forgot your wireless card at home. Test devices
_cannot_ be created unless NM is started with --enable-test-devices.
* src/NetworkManagerDbus.[ch]
- New "getLinkActive" method for devices
- New "setLinkActive" method for devices (only works on test devices)
- New "createTestDevice" method on NetworkManager object to create a test
device of a specified type (ie wired, wireless). UDI is created from
scratch, as is the interface name. Only works when NM is started with
--enable-test-devices switch.
- New "removeTestDevice" method on NetworkManager object which removes a
test device. Only works when NM is started with --enable-test-devices
* src/NetworkManagerDevice.[ch]
- Logic to facilitate test devices. Add variables to NMDevice struct to indicate
whether a device is a test device or not, and what its link status is.
- Deal with test devices in most functions. For those that work directly on hardware
special-case test devices.
- (nm_device_new): don't create a test device if test devices weren't enabled on the
command-line.
- (nm_device_update_link_active): split out logic for wired and wireless device link
checking to separate functions to facilitate test device link checking.
- (nm_device_set_enc_key): Since some drivers for wireless cards are daft and
don't make a distinction between System Authentication and Encryption
(namely Cisco aironet), we use Open System auth when setting a WEP key
on the card. We don't deal with Shared Key auth yet.
- (nm_device_activation_worker): split the activation cancel check logic out into
a separate routine nm_device_activation_cancel_if_needed()
- (nm_device_activation_signal_cancel): rename from nm_device_activation_cancel()
- (nm_device_fake_ap_list): Test wireless devices obviously cannot scan, so create
a list of fake access points that they can "see"
- (nm_device_is_test_device): return whether or not a device is a test device
* src/NetworkManagerPolicy.c
- (nm_policy_get_best_device): attempt to deal with wireless network selection,
previously if you "locked"/forced NM to use a wireless device but then
selected a wireless network for NM to use, it would switch to a wired device.
So, if the active device is wireless and it has a "forced" best AP, use it
if the "forced" best AP is still valid
- (nm_state_modification_monitor): deal with NULL best devices, for example
there were no usable network devices, or the last one was removed
* src/backends/NetworkManager*.c
- Deal with test devices, mostly just return success for operations like getting
a DHCP address
* test/nmtestdevices.c
- Test tool to create/remove/link-switch test devices
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@112 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-31 16:09:15 +00:00
|
|
|
struct iwreq wrq;
|
|
|
|
|
int iwlib_socket;
|
|
|
|
|
gboolean link = FALSE;
|
2004-06-24 14:18:37 +00:00
|
|
|
|
2004-08-31 Dan Williams <dcbw@redhat.com>
* Remove 'debug' extern global from all files since we now
use syslog()
* src/NetworkManager.[ch]
- Break out routine that get the net.interface property from HAL,
removing that logic from nm_create_device_and_add_to_list()
- (nm_create_device_and_add_to_list): make this a bit more general so
it doesn't do the talking to HAL. Also add arguments to facilitate
the create of test devices.
- (nm_data_mark_state_changed): rename from nm_data_set_state_modified()
- (nm_data_new, main, nm_print_usage): add new argument "--enable-test-devices"
which makes NetworkManager listen for dbus commands to create test
devices, which have no backing hardware. Use when you're on a plane
for example, and/or forgot your wireless card at home. Test devices
_cannot_ be created unless NM is started with --enable-test-devices.
* src/NetworkManagerDbus.[ch]
- New "getLinkActive" method for devices
- New "setLinkActive" method for devices (only works on test devices)
- New "createTestDevice" method on NetworkManager object to create a test
device of a specified type (ie wired, wireless). UDI is created from
scratch, as is the interface name. Only works when NM is started with
--enable-test-devices switch.
- New "removeTestDevice" method on NetworkManager object which removes a
test device. Only works when NM is started with --enable-test-devices
* src/NetworkManagerDevice.[ch]
- Logic to facilitate test devices. Add variables to NMDevice struct to indicate
whether a device is a test device or not, and what its link status is.
- Deal with test devices in most functions. For those that work directly on hardware
special-case test devices.
- (nm_device_new): don't create a test device if test devices weren't enabled on the
command-line.
- (nm_device_update_link_active): split out logic for wired and wireless device link
checking to separate functions to facilitate test device link checking.
- (nm_device_set_enc_key): Since some drivers for wireless cards are daft and
don't make a distinction between System Authentication and Encryption
(namely Cisco aironet), we use Open System auth when setting a WEP key
on the card. We don't deal with Shared Key auth yet.
- (nm_device_activation_worker): split the activation cancel check logic out into
a separate routine nm_device_activation_cancel_if_needed()
- (nm_device_activation_signal_cancel): rename from nm_device_activation_cancel()
- (nm_device_fake_ap_list): Test wireless devices obviously cannot scan, so create
a list of fake access points that they can "see"
- (nm_device_is_test_device): return whether or not a device is a test device
* src/NetworkManagerPolicy.c
- (nm_policy_get_best_device): attempt to deal with wireless network selection,
previously if you "locked"/forced NM to use a wireless device but then
selected a wireless network for NM to use, it would switch to a wired device.
So, if the active device is wireless and it has a "forced" best AP, use it
if the "forced" best AP is still valid
- (nm_state_modification_monitor): deal with NULL best devices, for example
there were no usable network devices, or the last one was removed
* src/backends/NetworkManager*.c
- Deal with test devices, mostly just return success for operations like getting
a DHCP address
* test/nmtestdevices.c
- Test tool to create/remove/link-switch test devices
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@112 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-31 16:09:15 +00:00
|
|
|
g_return_val_if_fail (dev != NULL, FALSE);
|
|
|
|
|
g_return_val_if_fail (dev->app_data != NULL, FALSE);
|
|
|
|
|
|
|
|
|
|
/* Test devices have their link state set through DBUS */
|
|
|
|
|
if (dev->test_device)
|
|
|
|
|
return (nm_device_get_link_active (dev));
|
2004-06-24 14:18:37 +00:00
|
|
|
|
2004-09-09 20:02:59 +00:00
|
|
|
/*
|
2004-06-24 14:18:37 +00:00
|
|
|
* For wireless cards, the best indicator of a "link" at this time
|
|
|
|
|
* seems to be whether the card has a valid access point MAC address.
|
|
|
|
|
* Is there a better way?
|
|
|
|
|
*/
|
2004-08-31 Dan Williams <dcbw@redhat.com>
* Remove 'debug' extern global from all files since we now
use syslog()
* src/NetworkManager.[ch]
- Break out routine that get the net.interface property from HAL,
removing that logic from nm_create_device_and_add_to_list()
- (nm_create_device_and_add_to_list): make this a bit more general so
it doesn't do the talking to HAL. Also add arguments to facilitate
the create of test devices.
- (nm_data_mark_state_changed): rename from nm_data_set_state_modified()
- (nm_data_new, main, nm_print_usage): add new argument "--enable-test-devices"
which makes NetworkManager listen for dbus commands to create test
devices, which have no backing hardware. Use when you're on a plane
for example, and/or forgot your wireless card at home. Test devices
_cannot_ be created unless NM is started with --enable-test-devices.
* src/NetworkManagerDbus.[ch]
- New "getLinkActive" method for devices
- New "setLinkActive" method for devices (only works on test devices)
- New "createTestDevice" method on NetworkManager object to create a test
device of a specified type (ie wired, wireless). UDI is created from
scratch, as is the interface name. Only works when NM is started with
--enable-test-devices switch.
- New "removeTestDevice" method on NetworkManager object which removes a
test device. Only works when NM is started with --enable-test-devices
* src/NetworkManagerDevice.[ch]
- Logic to facilitate test devices. Add variables to NMDevice struct to indicate
whether a device is a test device or not, and what its link status is.
- Deal with test devices in most functions. For those that work directly on hardware
special-case test devices.
- (nm_device_new): don't create a test device if test devices weren't enabled on the
command-line.
- (nm_device_update_link_active): split out logic for wired and wireless device link
checking to separate functions to facilitate test device link checking.
- (nm_device_set_enc_key): Since some drivers for wireless cards are daft and
don't make a distinction between System Authentication and Encryption
(namely Cisco aironet), we use Open System auth when setting a WEP key
on the card. We don't deal with Shared Key auth yet.
- (nm_device_activation_worker): split the activation cancel check logic out into
a separate routine nm_device_activation_cancel_if_needed()
- (nm_device_activation_signal_cancel): rename from nm_device_activation_cancel()
- (nm_device_fake_ap_list): Test wireless devices obviously cannot scan, so create
a list of fake access points that they can "see"
- (nm_device_is_test_device): return whether or not a device is a test device
* src/NetworkManagerPolicy.c
- (nm_policy_get_best_device): attempt to deal with wireless network selection,
previously if you "locked"/forced NM to use a wireless device but then
selected a wireless network for NM to use, it would switch to a wired device.
So, if the active device is wireless and it has a "forced" best AP, use it
if the "forced" best AP is still valid
- (nm_state_modification_monitor): deal with NULL best devices, for example
there were no usable network devices, or the last one was removed
* src/backends/NetworkManager*.c
- Deal with test devices, mostly just return success for operations like getting
a DHCP address
* test/nmtestdevices.c
- Test tool to create/remove/link-switch test devices
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@112 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-31 16:09:15 +00:00
|
|
|
iwlib_socket = iw_sockets_open ();
|
|
|
|
|
if (iw_get_ext (iwlib_socket, nm_device_get_iface (dev), SIOCGIWAP, &wrq) >= 0)
|
|
|
|
|
{
|
2004-10-29 16:37:43 +00:00
|
|
|
NMAccessPoint *best_ap;
|
|
|
|
|
|
|
|
|
|
if ((best_ap = nm_device_get_best_ap (dev)))
|
|
|
|
|
{
|
|
|
|
|
if ( nm_ethernet_address_is_valid ((struct ether_addr *)(&(wrq.u.ap_addr.sa_data)))
|
|
|
|
|
&& !nm_device_need_ap_switch (dev))
|
|
|
|
|
link = TRUE;
|
|
|
|
|
nm_ap_unref (best_ap);
|
|
|
|
|
}
|
2004-08-31 Dan Williams <dcbw@redhat.com>
* Remove 'debug' extern global from all files since we now
use syslog()
* src/NetworkManager.[ch]
- Break out routine that get the net.interface property from HAL,
removing that logic from nm_create_device_and_add_to_list()
- (nm_create_device_and_add_to_list): make this a bit more general so
it doesn't do the talking to HAL. Also add arguments to facilitate
the create of test devices.
- (nm_data_mark_state_changed): rename from nm_data_set_state_modified()
- (nm_data_new, main, nm_print_usage): add new argument "--enable-test-devices"
which makes NetworkManager listen for dbus commands to create test
devices, which have no backing hardware. Use when you're on a plane
for example, and/or forgot your wireless card at home. Test devices
_cannot_ be created unless NM is started with --enable-test-devices.
* src/NetworkManagerDbus.[ch]
- New "getLinkActive" method for devices
- New "setLinkActive" method for devices (only works on test devices)
- New "createTestDevice" method on NetworkManager object to create a test
device of a specified type (ie wired, wireless). UDI is created from
scratch, as is the interface name. Only works when NM is started with
--enable-test-devices switch.
- New "removeTestDevice" method on NetworkManager object which removes a
test device. Only works when NM is started with --enable-test-devices
* src/NetworkManagerDevice.[ch]
- Logic to facilitate test devices. Add variables to NMDevice struct to indicate
whether a device is a test device or not, and what its link status is.
- Deal with test devices in most functions. For those that work directly on hardware
special-case test devices.
- (nm_device_new): don't create a test device if test devices weren't enabled on the
command-line.
- (nm_device_update_link_active): split out logic for wired and wireless device link
checking to separate functions to facilitate test device link checking.
- (nm_device_set_enc_key): Since some drivers for wireless cards are daft and
don't make a distinction between System Authentication and Encryption
(namely Cisco aironet), we use Open System auth when setting a WEP key
on the card. We don't deal with Shared Key auth yet.
- (nm_device_activation_worker): split the activation cancel check logic out into
a separate routine nm_device_activation_cancel_if_needed()
- (nm_device_activation_signal_cancel): rename from nm_device_activation_cancel()
- (nm_device_fake_ap_list): Test wireless devices obviously cannot scan, so create
a list of fake access points that they can "see"
- (nm_device_is_test_device): return whether or not a device is a test device
* src/NetworkManagerPolicy.c
- (nm_policy_get_best_device): attempt to deal with wireless network selection,
previously if you "locked"/forced NM to use a wireless device but then
selected a wireless network for NM to use, it would switch to a wired device.
So, if the active device is wireless and it has a "forced" best AP, use it
if the "forced" best AP is still valid
- (nm_state_modification_monitor): deal with NULL best devices, for example
there were no usable network devices, or the last one was removed
* src/backends/NetworkManager*.c
- Deal with test devices, mostly just return success for operations like getting
a DHCP address
* test/nmtestdevices.c
- Test tool to create/remove/link-switch test devices
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@112 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-31 16:09:15 +00:00
|
|
|
}
|
|
|
|
|
close (iwlib_socket);
|
|
|
|
|
|
|
|
|
|
return (link);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* nm_device_wired_link_active
|
|
|
|
|
*
|
|
|
|
|
* Return the link state of a wired device. We usually just grab the HAL
|
2004-09-22 17:34:15 +00:00
|
|
|
* net.80203.link property, but on card insertion we need to check the MII
|
2004-08-31 Dan Williams <dcbw@redhat.com>
* Remove 'debug' extern global from all files since we now
use syslog()
* src/NetworkManager.[ch]
- Break out routine that get the net.interface property from HAL,
removing that logic from nm_create_device_and_add_to_list()
- (nm_create_device_and_add_to_list): make this a bit more general so
it doesn't do the talking to HAL. Also add arguments to facilitate
the create of test devices.
- (nm_data_mark_state_changed): rename from nm_data_set_state_modified()
- (nm_data_new, main, nm_print_usage): add new argument "--enable-test-devices"
which makes NetworkManager listen for dbus commands to create test
devices, which have no backing hardware. Use when you're on a plane
for example, and/or forgot your wireless card at home. Test devices
_cannot_ be created unless NM is started with --enable-test-devices.
* src/NetworkManagerDbus.[ch]
- New "getLinkActive" method for devices
- New "setLinkActive" method for devices (only works on test devices)
- New "createTestDevice" method on NetworkManager object to create a test
device of a specified type (ie wired, wireless). UDI is created from
scratch, as is the interface name. Only works when NM is started with
--enable-test-devices switch.
- New "removeTestDevice" method on NetworkManager object which removes a
test device. Only works when NM is started with --enable-test-devices
* src/NetworkManagerDevice.[ch]
- Logic to facilitate test devices. Add variables to NMDevice struct to indicate
whether a device is a test device or not, and what its link status is.
- Deal with test devices in most functions. For those that work directly on hardware
special-case test devices.
- (nm_device_new): don't create a test device if test devices weren't enabled on the
command-line.
- (nm_device_update_link_active): split out logic for wired and wireless device link
checking to separate functions to facilitate test device link checking.
- (nm_device_set_enc_key): Since some drivers for wireless cards are daft and
don't make a distinction between System Authentication and Encryption
(namely Cisco aironet), we use Open System auth when setting a WEP key
on the card. We don't deal with Shared Key auth yet.
- (nm_device_activation_worker): split the activation cancel check logic out into
a separate routine nm_device_activation_cancel_if_needed()
- (nm_device_activation_signal_cancel): rename from nm_device_activation_cancel()
- (nm_device_fake_ap_list): Test wireless devices obviously cannot scan, so create
a list of fake access points that they can "see"
- (nm_device_is_test_device): return whether or not a device is a test device
* src/NetworkManagerPolicy.c
- (nm_policy_get_best_device): attempt to deal with wireless network selection,
previously if you "locked"/forced NM to use a wireless device but then
selected a wireless network for NM to use, it would switch to a wired device.
So, if the active device is wireless and it has a "forced" best AP, use it
if the "forced" best AP is still valid
- (nm_state_modification_monitor): deal with NULL best devices, for example
there were no usable network devices, or the last one was removed
* src/backends/NetworkManager*.c
- Deal with test devices, mostly just return success for operations like getting
a DHCP address
* test/nmtestdevices.c
- Test tool to create/remove/link-switch test devices
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@112 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-31 16:09:15 +00:00
|
|
|
* registers of the card to get a more accurate response, since HAL may not
|
|
|
|
|
* have received a netlink socket link event for the device yet, and therefore
|
|
|
|
|
* will return FALSE when the device really does have a link.
|
|
|
|
|
*
|
|
|
|
|
*/
|
|
|
|
|
static gboolean nm_device_wired_link_active (NMDevice *dev, gboolean check_mii)
|
|
|
|
|
{
|
|
|
|
|
gboolean link = FALSE;
|
|
|
|
|
|
|
|
|
|
g_return_val_if_fail (dev != NULL, FALSE);
|
|
|
|
|
g_return_val_if_fail (dev->app_data != NULL, FALSE);
|
|
|
|
|
|
|
|
|
|
/* Test devices have their link state set through DBUS */
|
|
|
|
|
if (dev->test_device)
|
|
|
|
|
return (nm_device_get_link_active (dev));
|
|
|
|
|
|
|
|
|
|
if (check_mii)
|
|
|
|
|
link = mii_get_link (dev);
|
2004-09-22 17:34:15 +00:00
|
|
|
else if (hal_device_property_exists (dev->app_data->hal_ctx, nm_device_get_udi (dev), "net.80203.link"))
|
|
|
|
|
link = hal_device_get_property_bool (dev->app_data->hal_ctx, nm_device_get_udi (dev), "net.80203.link");
|
2004-08-31 Dan Williams <dcbw@redhat.com>
* Remove 'debug' extern global from all files since we now
use syslog()
* src/NetworkManager.[ch]
- Break out routine that get the net.interface property from HAL,
removing that logic from nm_create_device_and_add_to_list()
- (nm_create_device_and_add_to_list): make this a bit more general so
it doesn't do the talking to HAL. Also add arguments to facilitate
the create of test devices.
- (nm_data_mark_state_changed): rename from nm_data_set_state_modified()
- (nm_data_new, main, nm_print_usage): add new argument "--enable-test-devices"
which makes NetworkManager listen for dbus commands to create test
devices, which have no backing hardware. Use when you're on a plane
for example, and/or forgot your wireless card at home. Test devices
_cannot_ be created unless NM is started with --enable-test-devices.
* src/NetworkManagerDbus.[ch]
- New "getLinkActive" method for devices
- New "setLinkActive" method for devices (only works on test devices)
- New "createTestDevice" method on NetworkManager object to create a test
device of a specified type (ie wired, wireless). UDI is created from
scratch, as is the interface name. Only works when NM is started with
--enable-test-devices switch.
- New "removeTestDevice" method on NetworkManager object which removes a
test device. Only works when NM is started with --enable-test-devices
* src/NetworkManagerDevice.[ch]
- Logic to facilitate test devices. Add variables to NMDevice struct to indicate
whether a device is a test device or not, and what its link status is.
- Deal with test devices in most functions. For those that work directly on hardware
special-case test devices.
- (nm_device_new): don't create a test device if test devices weren't enabled on the
command-line.
- (nm_device_update_link_active): split out logic for wired and wireless device link
checking to separate functions to facilitate test device link checking.
- (nm_device_set_enc_key): Since some drivers for wireless cards are daft and
don't make a distinction between System Authentication and Encryption
(namely Cisco aironet), we use Open System auth when setting a WEP key
on the card. We don't deal with Shared Key auth yet.
- (nm_device_activation_worker): split the activation cancel check logic out into
a separate routine nm_device_activation_cancel_if_needed()
- (nm_device_activation_signal_cancel): rename from nm_device_activation_cancel()
- (nm_device_fake_ap_list): Test wireless devices obviously cannot scan, so create
a list of fake access points that they can "see"
- (nm_device_is_test_device): return whether or not a device is a test device
* src/NetworkManagerPolicy.c
- (nm_policy_get_best_device): attempt to deal with wireless network selection,
previously if you "locked"/forced NM to use a wireless device but then
selected a wireless network for NM to use, it would switch to a wired device.
So, if the active device is wireless and it has a "forced" best AP, use it
if the "forced" best AP is still valid
- (nm_state_modification_monitor): deal with NULL best devices, for example
there were no usable network devices, or the last one was removed
* src/backends/NetworkManager*.c
- Deal with test devices, mostly just return success for operations like getting
a DHCP address
* test/nmtestdevices.c
- Test tool to create/remove/link-switch test devices
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@112 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-31 16:09:15 +00:00
|
|
|
|
|
|
|
|
return (link);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* nm_device_update_link_active
|
|
|
|
|
*
|
|
|
|
|
* Updates the link state for a particular device.
|
|
|
|
|
*
|
|
|
|
|
*/
|
|
|
|
|
void nm_device_update_link_active (NMDevice *dev, gboolean check_mii)
|
|
|
|
|
{
|
|
|
|
|
gboolean link = FALSE;
|
|
|
|
|
|
|
|
|
|
g_return_if_fail (dev != NULL);
|
|
|
|
|
g_return_if_fail (dev->app_data != NULL);
|
|
|
|
|
|
2004-08-05 18:54:29 +00:00
|
|
|
switch (nm_device_get_type (dev))
|
2004-06-24 14:18:37 +00:00
|
|
|
{
|
2004-08-05 18:54:29 +00:00
|
|
|
case DEVICE_TYPE_WIRELESS_ETHERNET:
|
2004-08-31 Dan Williams <dcbw@redhat.com>
* Remove 'debug' extern global from all files since we now
use syslog()
* src/NetworkManager.[ch]
- Break out routine that get the net.interface property from HAL,
removing that logic from nm_create_device_and_add_to_list()
- (nm_create_device_and_add_to_list): make this a bit more general so
it doesn't do the talking to HAL. Also add arguments to facilitate
the create of test devices.
- (nm_data_mark_state_changed): rename from nm_data_set_state_modified()
- (nm_data_new, main, nm_print_usage): add new argument "--enable-test-devices"
which makes NetworkManager listen for dbus commands to create test
devices, which have no backing hardware. Use when you're on a plane
for example, and/or forgot your wireless card at home. Test devices
_cannot_ be created unless NM is started with --enable-test-devices.
* src/NetworkManagerDbus.[ch]
- New "getLinkActive" method for devices
- New "setLinkActive" method for devices (only works on test devices)
- New "createTestDevice" method on NetworkManager object to create a test
device of a specified type (ie wired, wireless). UDI is created from
scratch, as is the interface name. Only works when NM is started with
--enable-test-devices switch.
- New "removeTestDevice" method on NetworkManager object which removes a
test device. Only works when NM is started with --enable-test-devices
* src/NetworkManagerDevice.[ch]
- Logic to facilitate test devices. Add variables to NMDevice struct to indicate
whether a device is a test device or not, and what its link status is.
- Deal with test devices in most functions. For those that work directly on hardware
special-case test devices.
- (nm_device_new): don't create a test device if test devices weren't enabled on the
command-line.
- (nm_device_update_link_active): split out logic for wired and wireless device link
checking to separate functions to facilitate test device link checking.
- (nm_device_set_enc_key): Since some drivers for wireless cards are daft and
don't make a distinction between System Authentication and Encryption
(namely Cisco aironet), we use Open System auth when setting a WEP key
on the card. We don't deal with Shared Key auth yet.
- (nm_device_activation_worker): split the activation cancel check logic out into
a separate routine nm_device_activation_cancel_if_needed()
- (nm_device_activation_signal_cancel): rename from nm_device_activation_cancel()
- (nm_device_fake_ap_list): Test wireless devices obviously cannot scan, so create
a list of fake access points that they can "see"
- (nm_device_is_test_device): return whether or not a device is a test device
* src/NetworkManagerPolicy.c
- (nm_policy_get_best_device): attempt to deal with wireless network selection,
previously if you "locked"/forced NM to use a wireless device but then
selected a wireless network for NM to use, it would switch to a wired device.
So, if the active device is wireless and it has a "forced" best AP, use it
if the "forced" best AP is still valid
- (nm_state_modification_monitor): deal with NULL best devices, for example
there were no usable network devices, or the last one was removed
* src/backends/NetworkManager*.c
- Deal with test devices, mostly just return success for operations like getting
a DHCP address
* test/nmtestdevices.c
- Test tool to create/remove/link-switch test devices
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@112 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-31 16:09:15 +00:00
|
|
|
link = nm_device_wireless_link_active (dev);
|
2004-09-13 17:43:16 +00:00
|
|
|
/* Update our current signal strength too */
|
|
|
|
|
nm_device_update_signal_strength (dev);
|
2004-06-24 14:18:37 +00:00
|
|
|
break;
|
|
|
|
|
|
2004-08-05 18:54:29 +00:00
|
|
|
case DEVICE_TYPE_WIRED_ETHERNET:
|
2004-08-31 Dan Williams <dcbw@redhat.com>
* Remove 'debug' extern global from all files since we now
use syslog()
* src/NetworkManager.[ch]
- Break out routine that get the net.interface property from HAL,
removing that logic from nm_create_device_and_add_to_list()
- (nm_create_device_and_add_to_list): make this a bit more general so
it doesn't do the talking to HAL. Also add arguments to facilitate
the create of test devices.
- (nm_data_mark_state_changed): rename from nm_data_set_state_modified()
- (nm_data_new, main, nm_print_usage): add new argument "--enable-test-devices"
which makes NetworkManager listen for dbus commands to create test
devices, which have no backing hardware. Use when you're on a plane
for example, and/or forgot your wireless card at home. Test devices
_cannot_ be created unless NM is started with --enable-test-devices.
* src/NetworkManagerDbus.[ch]
- New "getLinkActive" method for devices
- New "setLinkActive" method for devices (only works on test devices)
- New "createTestDevice" method on NetworkManager object to create a test
device of a specified type (ie wired, wireless). UDI is created from
scratch, as is the interface name. Only works when NM is started with
--enable-test-devices switch.
- New "removeTestDevice" method on NetworkManager object which removes a
test device. Only works when NM is started with --enable-test-devices
* src/NetworkManagerDevice.[ch]
- Logic to facilitate test devices. Add variables to NMDevice struct to indicate
whether a device is a test device or not, and what its link status is.
- Deal with test devices in most functions. For those that work directly on hardware
special-case test devices.
- (nm_device_new): don't create a test device if test devices weren't enabled on the
command-line.
- (nm_device_update_link_active): split out logic for wired and wireless device link
checking to separate functions to facilitate test device link checking.
- (nm_device_set_enc_key): Since some drivers for wireless cards are daft and
don't make a distinction between System Authentication and Encryption
(namely Cisco aironet), we use Open System auth when setting a WEP key
on the card. We don't deal with Shared Key auth yet.
- (nm_device_activation_worker): split the activation cancel check logic out into
a separate routine nm_device_activation_cancel_if_needed()
- (nm_device_activation_signal_cancel): rename from nm_device_activation_cancel()
- (nm_device_fake_ap_list): Test wireless devices obviously cannot scan, so create
a list of fake access points that they can "see"
- (nm_device_is_test_device): return whether or not a device is a test device
* src/NetworkManagerPolicy.c
- (nm_policy_get_best_device): attempt to deal with wireless network selection,
previously if you "locked"/forced NM to use a wireless device but then
selected a wireless network for NM to use, it would switch to a wired device.
So, if the active device is wireless and it has a "forced" best AP, use it
if the "forced" best AP is still valid
- (nm_state_modification_monitor): deal with NULL best devices, for example
there were no usable network devices, or the last one was removed
* src/backends/NetworkManager*.c
- Deal with test devices, mostly just return success for operations like getting
a DHCP address
* test/nmtestdevices.c
- Test tool to create/remove/link-switch test devices
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@112 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-31 16:09:15 +00:00
|
|
|
link = nm_device_wired_link_active (dev, check_mii);
|
2004-09-08 18:14:42 +00:00
|
|
|
break;
|
2004-06-24 14:18:37 +00:00
|
|
|
|
|
|
|
|
default:
|
2004-08-31 Dan Williams <dcbw@redhat.com>
* Remove 'debug' extern global from all files since we now
use syslog()
* src/NetworkManager.[ch]
- Break out routine that get the net.interface property from HAL,
removing that logic from nm_create_device_and_add_to_list()
- (nm_create_device_and_add_to_list): make this a bit more general so
it doesn't do the talking to HAL. Also add arguments to facilitate
the create of test devices.
- (nm_data_mark_state_changed): rename from nm_data_set_state_modified()
- (nm_data_new, main, nm_print_usage): add new argument "--enable-test-devices"
which makes NetworkManager listen for dbus commands to create test
devices, which have no backing hardware. Use when you're on a plane
for example, and/or forgot your wireless card at home. Test devices
_cannot_ be created unless NM is started with --enable-test-devices.
* src/NetworkManagerDbus.[ch]
- New "getLinkActive" method for devices
- New "setLinkActive" method for devices (only works on test devices)
- New "createTestDevice" method on NetworkManager object to create a test
device of a specified type (ie wired, wireless). UDI is created from
scratch, as is the interface name. Only works when NM is started with
--enable-test-devices switch.
- New "removeTestDevice" method on NetworkManager object which removes a
test device. Only works when NM is started with --enable-test-devices
* src/NetworkManagerDevice.[ch]
- Logic to facilitate test devices. Add variables to NMDevice struct to indicate
whether a device is a test device or not, and what its link status is.
- Deal with test devices in most functions. For those that work directly on hardware
special-case test devices.
- (nm_device_new): don't create a test device if test devices weren't enabled on the
command-line.
- (nm_device_update_link_active): split out logic for wired and wireless device link
checking to separate functions to facilitate test device link checking.
- (nm_device_set_enc_key): Since some drivers for wireless cards are daft and
don't make a distinction between System Authentication and Encryption
(namely Cisco aironet), we use Open System auth when setting a WEP key
on the card. We don't deal with Shared Key auth yet.
- (nm_device_activation_worker): split the activation cancel check logic out into
a separate routine nm_device_activation_cancel_if_needed()
- (nm_device_activation_signal_cancel): rename from nm_device_activation_cancel()
- (nm_device_fake_ap_list): Test wireless devices obviously cannot scan, so create
a list of fake access points that they can "see"
- (nm_device_is_test_device): return whether or not a device is a test device
* src/NetworkManagerPolicy.c
- (nm_policy_get_best_device): attempt to deal with wireless network selection,
previously if you "locked"/forced NM to use a wireless device but then
selected a wireless network for NM to use, it would switch to a wired device.
So, if the active device is wireless and it has a "forced" best AP, use it
if the "forced" best AP is still valid
- (nm_state_modification_monitor): deal with NULL best devices, for example
there were no usable network devices, or the last one was removed
* src/backends/NetworkManager*.c
- Deal with test devices, mostly just return success for operations like getting
a DHCP address
* test/nmtestdevices.c
- Test tool to create/remove/link-switch test devices
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@112 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-31 16:09:15 +00:00
|
|
|
link = nm_device_get_link_active (dev); /* Can't get link info for this device, so don't change link status */
|
2004-06-24 14:18:37 +00:00
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* Update device link status and global state variable if the status changed */
|
2004-08-31 Dan Williams <dcbw@redhat.com>
* Remove 'debug' extern global from all files since we now
use syslog()
* src/NetworkManager.[ch]
- Break out routine that get the net.interface property from HAL,
removing that logic from nm_create_device_and_add_to_list()
- (nm_create_device_and_add_to_list): make this a bit more general so
it doesn't do the talking to HAL. Also add arguments to facilitate
the create of test devices.
- (nm_data_mark_state_changed): rename from nm_data_set_state_modified()
- (nm_data_new, main, nm_print_usage): add new argument "--enable-test-devices"
which makes NetworkManager listen for dbus commands to create test
devices, which have no backing hardware. Use when you're on a plane
for example, and/or forgot your wireless card at home. Test devices
_cannot_ be created unless NM is started with --enable-test-devices.
* src/NetworkManagerDbus.[ch]
- New "getLinkActive" method for devices
- New "setLinkActive" method for devices (only works on test devices)
- New "createTestDevice" method on NetworkManager object to create a test
device of a specified type (ie wired, wireless). UDI is created from
scratch, as is the interface name. Only works when NM is started with
--enable-test-devices switch.
- New "removeTestDevice" method on NetworkManager object which removes a
test device. Only works when NM is started with --enable-test-devices
* src/NetworkManagerDevice.[ch]
- Logic to facilitate test devices. Add variables to NMDevice struct to indicate
whether a device is a test device or not, and what its link status is.
- Deal with test devices in most functions. For those that work directly on hardware
special-case test devices.
- (nm_device_new): don't create a test device if test devices weren't enabled on the
command-line.
- (nm_device_update_link_active): split out logic for wired and wireless device link
checking to separate functions to facilitate test device link checking.
- (nm_device_set_enc_key): Since some drivers for wireless cards are daft and
don't make a distinction between System Authentication and Encryption
(namely Cisco aironet), we use Open System auth when setting a WEP key
on the card. We don't deal with Shared Key auth yet.
- (nm_device_activation_worker): split the activation cancel check logic out into
a separate routine nm_device_activation_cancel_if_needed()
- (nm_device_activation_signal_cancel): rename from nm_device_activation_cancel()
- (nm_device_fake_ap_list): Test wireless devices obviously cannot scan, so create
a list of fake access points that they can "see"
- (nm_device_is_test_device): return whether or not a device is a test device
* src/NetworkManagerPolicy.c
- (nm_policy_get_best_device): attempt to deal with wireless network selection,
previously if you "locked"/forced NM to use a wireless device but then
selected a wireless network for NM to use, it would switch to a wired device.
So, if the active device is wireless and it has a "forced" best AP, use it
if the "forced" best AP is still valid
- (nm_state_modification_monitor): deal with NULL best devices, for example
there were no usable network devices, or the last one was removed
* src/backends/NetworkManager*.c
- Deal with test devices, mostly just return success for operations like getting
a DHCP address
* test/nmtestdevices.c
- Test tool to create/remove/link-switch test devices
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@112 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-31 16:09:15 +00:00
|
|
|
if (link != nm_device_get_link_active (dev))
|
2004-06-24 14:18:37 +00:00
|
|
|
{
|
2004-08-31 Dan Williams <dcbw@redhat.com>
* Remove 'debug' extern global from all files since we now
use syslog()
* src/NetworkManager.[ch]
- Break out routine that get the net.interface property from HAL,
removing that logic from nm_create_device_and_add_to_list()
- (nm_create_device_and_add_to_list): make this a bit more general so
it doesn't do the talking to HAL. Also add arguments to facilitate
the create of test devices.
- (nm_data_mark_state_changed): rename from nm_data_set_state_modified()
- (nm_data_new, main, nm_print_usage): add new argument "--enable-test-devices"
which makes NetworkManager listen for dbus commands to create test
devices, which have no backing hardware. Use when you're on a plane
for example, and/or forgot your wireless card at home. Test devices
_cannot_ be created unless NM is started with --enable-test-devices.
* src/NetworkManagerDbus.[ch]
- New "getLinkActive" method for devices
- New "setLinkActive" method for devices (only works on test devices)
- New "createTestDevice" method on NetworkManager object to create a test
device of a specified type (ie wired, wireless). UDI is created from
scratch, as is the interface name. Only works when NM is started with
--enable-test-devices switch.
- New "removeTestDevice" method on NetworkManager object which removes a
test device. Only works when NM is started with --enable-test-devices
* src/NetworkManagerDevice.[ch]
- Logic to facilitate test devices. Add variables to NMDevice struct to indicate
whether a device is a test device or not, and what its link status is.
- Deal with test devices in most functions. For those that work directly on hardware
special-case test devices.
- (nm_device_new): don't create a test device if test devices weren't enabled on the
command-line.
- (nm_device_update_link_active): split out logic for wired and wireless device link
checking to separate functions to facilitate test device link checking.
- (nm_device_set_enc_key): Since some drivers for wireless cards are daft and
don't make a distinction between System Authentication and Encryption
(namely Cisco aironet), we use Open System auth when setting a WEP key
on the card. We don't deal with Shared Key auth yet.
- (nm_device_activation_worker): split the activation cancel check logic out into
a separate routine nm_device_activation_cancel_if_needed()
- (nm_device_activation_signal_cancel): rename from nm_device_activation_cancel()
- (nm_device_fake_ap_list): Test wireless devices obviously cannot scan, so create
a list of fake access points that they can "see"
- (nm_device_is_test_device): return whether or not a device is a test device
* src/NetworkManagerPolicy.c
- (nm_policy_get_best_device): attempt to deal with wireless network selection,
previously if you "locked"/forced NM to use a wireless device but then
selected a wireless network for NM to use, it would switch to a wired device.
So, if the active device is wireless and it has a "forced" best AP, use it
if the "forced" best AP is still valid
- (nm_state_modification_monitor): deal with NULL best devices, for example
there were no usable network devices, or the last one was removed
* src/backends/NetworkManager*.c
- Deal with test devices, mostly just return success for operations like getting
a DHCP address
* test/nmtestdevices.c
- Test tool to create/remove/link-switch test devices
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@112 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-31 16:09:15 +00:00
|
|
|
nm_device_set_link_active (dev, link);
|
|
|
|
|
nm_data_mark_state_changed (dev->app_data);
|
2004-06-24 14:18:37 +00:00
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* nm_device_get_essid
|
|
|
|
|
*
|
|
|
|
|
* If a device is wireless, return the essid that it is attempting
|
|
|
|
|
* to use.
|
|
|
|
|
*
|
|
|
|
|
* Returns: allocated string containing essid. Must be freed by caller.
|
|
|
|
|
*
|
|
|
|
|
*/
|
|
|
|
|
char * nm_device_get_essid (NMDevice *dev)
|
|
|
|
|
{
|
|
|
|
|
int iwlib_socket;
|
|
|
|
|
int err;
|
|
|
|
|
|
|
|
|
|
g_return_val_if_fail (dev != NULL, NULL);
|
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
|
|
|
g_return_val_if_fail (nm_device_is_wireless (dev), NULL);
|
2004-08-31 Dan Williams <dcbw@redhat.com>
* Remove 'debug' extern global from all files since we now
use syslog()
* src/NetworkManager.[ch]
- Break out routine that get the net.interface property from HAL,
removing that logic from nm_create_device_and_add_to_list()
- (nm_create_device_and_add_to_list): make this a bit more general so
it doesn't do the talking to HAL. Also add arguments to facilitate
the create of test devices.
- (nm_data_mark_state_changed): rename from nm_data_set_state_modified()
- (nm_data_new, main, nm_print_usage): add new argument "--enable-test-devices"
which makes NetworkManager listen for dbus commands to create test
devices, which have no backing hardware. Use when you're on a plane
for example, and/or forgot your wireless card at home. Test devices
_cannot_ be created unless NM is started with --enable-test-devices.
* src/NetworkManagerDbus.[ch]
- New "getLinkActive" method for devices
- New "setLinkActive" method for devices (only works on test devices)
- New "createTestDevice" method on NetworkManager object to create a test
device of a specified type (ie wired, wireless). UDI is created from
scratch, as is the interface name. Only works when NM is started with
--enable-test-devices switch.
- New "removeTestDevice" method on NetworkManager object which removes a
test device. Only works when NM is started with --enable-test-devices
* src/NetworkManagerDevice.[ch]
- Logic to facilitate test devices. Add variables to NMDevice struct to indicate
whether a device is a test device or not, and what its link status is.
- Deal with test devices in most functions. For those that work directly on hardware
special-case test devices.
- (nm_device_new): don't create a test device if test devices weren't enabled on the
command-line.
- (nm_device_update_link_active): split out logic for wired and wireless device link
checking to separate functions to facilitate test device link checking.
- (nm_device_set_enc_key): Since some drivers for wireless cards are daft and
don't make a distinction between System Authentication and Encryption
(namely Cisco aironet), we use Open System auth when setting a WEP key
on the card. We don't deal with Shared Key auth yet.
- (nm_device_activation_worker): split the activation cancel check logic out into
a separate routine nm_device_activation_cancel_if_needed()
- (nm_device_activation_signal_cancel): rename from nm_device_activation_cancel()
- (nm_device_fake_ap_list): Test wireless devices obviously cannot scan, so create
a list of fake access points that they can "see"
- (nm_device_is_test_device): return whether or not a device is a test device
* src/NetworkManagerPolicy.c
- (nm_policy_get_best_device): attempt to deal with wireless network selection,
previously if you "locked"/forced NM to use a wireless device but then
selected a wireless network for NM to use, it would switch to a wired device.
So, if the active device is wireless and it has a "forced" best AP, use it
if the "forced" best AP is still valid
- (nm_state_modification_monitor): deal with NULL best devices, for example
there were no usable network devices, or the last one was removed
* src/backends/NetworkManager*.c
- Deal with test devices, mostly just return success for operations like getting
a DHCP address
* test/nmtestdevices.c
- Test tool to create/remove/link-switch test devices
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@112 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-31 16:09:15 +00:00
|
|
|
|
|
|
|
|
/* Test devices return the essid of their "best" access point
|
|
|
|
|
* or if there is none, the contents of the cur_essid field.
|
|
|
|
|
*/
|
|
|
|
|
if (dev->test_device)
|
|
|
|
|
{
|
2004-10-29 16:37:43 +00:00
|
|
|
NMAccessPoint *best_ap = nm_device_get_best_ap (dev);
|
|
|
|
|
char *essid = dev->options.wireless.cur_essid;
|
|
|
|
|
|
|
|
|
|
/* Or, if we've got a best ap, use that ESSID instead */
|
|
|
|
|
if (best_ap)
|
|
|
|
|
{
|
|
|
|
|
essid = nm_ap_get_essid (best_ap);
|
|
|
|
|
nm_ap_unref (best_ap);
|
|
|
|
|
}
|
|
|
|
|
return (essid);
|
2004-08-31 Dan Williams <dcbw@redhat.com>
* Remove 'debug' extern global from all files since we now
use syslog()
* src/NetworkManager.[ch]
- Break out routine that get the net.interface property from HAL,
removing that logic from nm_create_device_and_add_to_list()
- (nm_create_device_and_add_to_list): make this a bit more general so
it doesn't do the talking to HAL. Also add arguments to facilitate
the create of test devices.
- (nm_data_mark_state_changed): rename from nm_data_set_state_modified()
- (nm_data_new, main, nm_print_usage): add new argument "--enable-test-devices"
which makes NetworkManager listen for dbus commands to create test
devices, which have no backing hardware. Use when you're on a plane
for example, and/or forgot your wireless card at home. Test devices
_cannot_ be created unless NM is started with --enable-test-devices.
* src/NetworkManagerDbus.[ch]
- New "getLinkActive" method for devices
- New "setLinkActive" method for devices (only works on test devices)
- New "createTestDevice" method on NetworkManager object to create a test
device of a specified type (ie wired, wireless). UDI is created from
scratch, as is the interface name. Only works when NM is started with
--enable-test-devices switch.
- New "removeTestDevice" method on NetworkManager object which removes a
test device. Only works when NM is started with --enable-test-devices
* src/NetworkManagerDevice.[ch]
- Logic to facilitate test devices. Add variables to NMDevice struct to indicate
whether a device is a test device or not, and what its link status is.
- Deal with test devices in most functions. For those that work directly on hardware
special-case test devices.
- (nm_device_new): don't create a test device if test devices weren't enabled on the
command-line.
- (nm_device_update_link_active): split out logic for wired and wireless device link
checking to separate functions to facilitate test device link checking.
- (nm_device_set_enc_key): Since some drivers for wireless cards are daft and
don't make a distinction between System Authentication and Encryption
(namely Cisco aironet), we use Open System auth when setting a WEP key
on the card. We don't deal with Shared Key auth yet.
- (nm_device_activation_worker): split the activation cancel check logic out into
a separate routine nm_device_activation_cancel_if_needed()
- (nm_device_activation_signal_cancel): rename from nm_device_activation_cancel()
- (nm_device_fake_ap_list): Test wireless devices obviously cannot scan, so create
a list of fake access points that they can "see"
- (nm_device_is_test_device): return whether or not a device is a test device
* src/NetworkManagerPolicy.c
- (nm_policy_get_best_device): attempt to deal with wireless network selection,
previously if you "locked"/forced NM to use a wireless device but then
selected a wireless network for NM to use, it would switch to a wired device.
So, if the active device is wireless and it has a "forced" best AP, use it
if the "forced" best AP is still valid
- (nm_state_modification_monitor): deal with NULL best devices, for example
there were no usable network devices, or the last one was removed
* src/backends/NetworkManager*.c
- Deal with test devices, mostly just return success for operations like getting
a DHCP address
* test/nmtestdevices.c
- Test tool to create/remove/link-switch test devices
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@112 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-31 16:09:15 +00:00
|
|
|
}
|
2004-06-24 14:18:37 +00:00
|
|
|
|
|
|
|
|
iwlib_socket = iw_sockets_open ();
|
|
|
|
|
if (iwlib_socket >= 0)
|
|
|
|
|
{
|
2004-09-13 21:00:35 +00:00
|
|
|
wireless_config info;
|
|
|
|
|
|
|
|
|
|
err = iw_get_basic_config(iwlib_socket, nm_device_get_iface (dev), &info);
|
|
|
|
|
if (err >= 0)
|
|
|
|
|
{
|
|
|
|
|
if (dev->options.wireless.cur_essid)
|
|
|
|
|
g_free (dev->options.wireless.cur_essid);
|
|
|
|
|
dev->options.wireless.cur_essid = g_strdup (info.essid);
|
|
|
|
|
}
|
|
|
|
|
else
|
|
|
|
|
syslog (LOG_ERR, "nm_device_get_essid(): error setting ESSID for device %s. errno = %d", nm_device_get_iface (dev), errno);
|
|
|
|
|
|
2004-06-24 14:18:37 +00:00
|
|
|
close (iwlib_socket);
|
|
|
|
|
}
|
|
|
|
|
|
2004-08-05 18:54:29 +00:00
|
|
|
return (dev->options.wireless.cur_essid);
|
2004-06-24 14:18:37 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* nm_device_set_essid
|
|
|
|
|
*
|
|
|
|
|
* If a device is wireless, set the essid that it should use.
|
|
|
|
|
*/
|
|
|
|
|
void nm_device_set_essid (NMDevice *dev, const char *essid)
|
|
|
|
|
{
|
|
|
|
|
int iwlib_socket;
|
|
|
|
|
int err;
|
|
|
|
|
struct iwreq wreq;
|
|
|
|
|
unsigned char safe_essid[IW_ESSID_MAX_SIZE + 1] = "\0";
|
|
|
|
|
|
|
|
|
|
g_return_if_fail (dev != NULL);
|
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
|
|
|
g_return_if_fail (nm_device_is_wireless (dev));
|
2004-06-24 14:18:37 +00:00
|
|
|
|
2004-08-31 Dan Williams <dcbw@redhat.com>
* Remove 'debug' extern global from all files since we now
use syslog()
* src/NetworkManager.[ch]
- Break out routine that get the net.interface property from HAL,
removing that logic from nm_create_device_and_add_to_list()
- (nm_create_device_and_add_to_list): make this a bit more general so
it doesn't do the talking to HAL. Also add arguments to facilitate
the create of test devices.
- (nm_data_mark_state_changed): rename from nm_data_set_state_modified()
- (nm_data_new, main, nm_print_usage): add new argument "--enable-test-devices"
which makes NetworkManager listen for dbus commands to create test
devices, which have no backing hardware. Use when you're on a plane
for example, and/or forgot your wireless card at home. Test devices
_cannot_ be created unless NM is started with --enable-test-devices.
* src/NetworkManagerDbus.[ch]
- New "getLinkActive" method for devices
- New "setLinkActive" method for devices (only works on test devices)
- New "createTestDevice" method on NetworkManager object to create a test
device of a specified type (ie wired, wireless). UDI is created from
scratch, as is the interface name. Only works when NM is started with
--enable-test-devices switch.
- New "removeTestDevice" method on NetworkManager object which removes a
test device. Only works when NM is started with --enable-test-devices
* src/NetworkManagerDevice.[ch]
- Logic to facilitate test devices. Add variables to NMDevice struct to indicate
whether a device is a test device or not, and what its link status is.
- Deal with test devices in most functions. For those that work directly on hardware
special-case test devices.
- (nm_device_new): don't create a test device if test devices weren't enabled on the
command-line.
- (nm_device_update_link_active): split out logic for wired and wireless device link
checking to separate functions to facilitate test device link checking.
- (nm_device_set_enc_key): Since some drivers for wireless cards are daft and
don't make a distinction between System Authentication and Encryption
(namely Cisco aironet), we use Open System auth when setting a WEP key
on the card. We don't deal with Shared Key auth yet.
- (nm_device_activation_worker): split the activation cancel check logic out into
a separate routine nm_device_activation_cancel_if_needed()
- (nm_device_activation_signal_cancel): rename from nm_device_activation_cancel()
- (nm_device_fake_ap_list): Test wireless devices obviously cannot scan, so create
a list of fake access points that they can "see"
- (nm_device_is_test_device): return whether or not a device is a test device
* src/NetworkManagerPolicy.c
- (nm_policy_get_best_device): attempt to deal with wireless network selection,
previously if you "locked"/forced NM to use a wireless device but then
selected a wireless network for NM to use, it would switch to a wired device.
So, if the active device is wireless and it has a "forced" best AP, use it
if the "forced" best AP is still valid
- (nm_state_modification_monitor): deal with NULL best devices, for example
there were no usable network devices, or the last one was removed
* src/backends/NetworkManager*.c
- Deal with test devices, mostly just return success for operations like getting
a DHCP address
* test/nmtestdevices.c
- Test tool to create/remove/link-switch test devices
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@112 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-31 16:09:15 +00:00
|
|
|
/* Test devices directly set cur_essid */
|
|
|
|
|
if (dev->test_device)
|
|
|
|
|
{
|
|
|
|
|
if (dev->options.wireless.cur_essid)
|
|
|
|
|
g_free (dev->options.wireless.cur_essid);
|
|
|
|
|
dev->options.wireless.cur_essid = g_strdup (essid);
|
|
|
|
|
return;
|
|
|
|
|
}
|
|
|
|
|
|
2004-06-24 14:18:37 +00:00
|
|
|
/* Make sure the essid we get passed is a valid size */
|
|
|
|
|
if (!essid)
|
|
|
|
|
safe_essid[0] = '\0';
|
|
|
|
|
else
|
|
|
|
|
{
|
|
|
|
|
strncpy (safe_essid, essid, IW_ESSID_MAX_SIZE);
|
|
|
|
|
safe_essid[IW_ESSID_MAX_SIZE] = '\0';
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
iwlib_socket = iw_sockets_open ();
|
|
|
|
|
if (iwlib_socket >= 0)
|
|
|
|
|
{
|
|
|
|
|
wreq.u.essid.pointer = (caddr_t) safe_essid;
|
|
|
|
|
wreq.u.essid.length = strlen (safe_essid) + 1;
|
|
|
|
|
wreq.u.essid.flags = 1; /* Enable essid on card */
|
|
|
|
|
|
|
|
|
|
err = iw_set_ext (iwlib_socket, nm_device_get_iface (dev), SIOCSIWESSID, &wreq);
|
|
|
|
|
if (err == -1)
|
2004-08-23 19:20:49 +00:00
|
|
|
syslog (LOG_ERR, "nm_device_set_essid(): error setting ESSID for device %s. errno = %d", nm_device_get_iface (dev), errno);
|
2004-06-24 14:18:37 +00:00
|
|
|
|
|
|
|
|
close (iwlib_socket);
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
2004-07-06 01:34:10 +00:00
|
|
|
/*
|
|
|
|
|
* nm_device_get_ap_address
|
|
|
|
|
*
|
|
|
|
|
* If a device is wireless, get the access point's ethernet address
|
|
|
|
|
* that the card is associated with.
|
|
|
|
|
*/
|
|
|
|
|
void nm_device_get_ap_address (NMDevice *dev, struct ether_addr *addr)
|
|
|
|
|
{
|
|
|
|
|
int iwlib_socket;
|
|
|
|
|
struct iwreq wrq;
|
|
|
|
|
|
|
|
|
|
g_return_if_fail (dev != NULL);
|
|
|
|
|
g_return_if_fail (addr != NULL);
|
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
|
|
|
g_return_if_fail (nm_device_is_wireless (dev));
|
2004-07-06 01:34:10 +00:00
|
|
|
|
2004-08-31 Dan Williams <dcbw@redhat.com>
* Remove 'debug' extern global from all files since we now
use syslog()
* src/NetworkManager.[ch]
- Break out routine that get the net.interface property from HAL,
removing that logic from nm_create_device_and_add_to_list()
- (nm_create_device_and_add_to_list): make this a bit more general so
it doesn't do the talking to HAL. Also add arguments to facilitate
the create of test devices.
- (nm_data_mark_state_changed): rename from nm_data_set_state_modified()
- (nm_data_new, main, nm_print_usage): add new argument "--enable-test-devices"
which makes NetworkManager listen for dbus commands to create test
devices, which have no backing hardware. Use when you're on a plane
for example, and/or forgot your wireless card at home. Test devices
_cannot_ be created unless NM is started with --enable-test-devices.
* src/NetworkManagerDbus.[ch]
- New "getLinkActive" method for devices
- New "setLinkActive" method for devices (only works on test devices)
- New "createTestDevice" method on NetworkManager object to create a test
device of a specified type (ie wired, wireless). UDI is created from
scratch, as is the interface name. Only works when NM is started with
--enable-test-devices switch.
- New "removeTestDevice" method on NetworkManager object which removes a
test device. Only works when NM is started with --enable-test-devices
* src/NetworkManagerDevice.[ch]
- Logic to facilitate test devices. Add variables to NMDevice struct to indicate
whether a device is a test device or not, and what its link status is.
- Deal with test devices in most functions. For those that work directly on hardware
special-case test devices.
- (nm_device_new): don't create a test device if test devices weren't enabled on the
command-line.
- (nm_device_update_link_active): split out logic for wired and wireless device link
checking to separate functions to facilitate test device link checking.
- (nm_device_set_enc_key): Since some drivers for wireless cards are daft and
don't make a distinction between System Authentication and Encryption
(namely Cisco aironet), we use Open System auth when setting a WEP key
on the card. We don't deal with Shared Key auth yet.
- (nm_device_activation_worker): split the activation cancel check logic out into
a separate routine nm_device_activation_cancel_if_needed()
- (nm_device_activation_signal_cancel): rename from nm_device_activation_cancel()
- (nm_device_fake_ap_list): Test wireless devices obviously cannot scan, so create
a list of fake access points that they can "see"
- (nm_device_is_test_device): return whether or not a device is a test device
* src/NetworkManagerPolicy.c
- (nm_policy_get_best_device): attempt to deal with wireless network selection,
previously if you "locked"/forced NM to use a wireless device but then
selected a wireless network for NM to use, it would switch to a wired device.
So, if the active device is wireless and it has a "forced" best AP, use it
if the "forced" best AP is still valid
- (nm_state_modification_monitor): deal with NULL best devices, for example
there were no usable network devices, or the last one was removed
* src/backends/NetworkManager*.c
- Deal with test devices, mostly just return success for operations like getting
a DHCP address
* test/nmtestdevices.c
- Test tool to create/remove/link-switch test devices
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@112 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-31 16:09:15 +00:00
|
|
|
/* Test devices return an invalid address when there's no link,
|
|
|
|
|
* and a made-up address when there is a link.
|
|
|
|
|
*/
|
|
|
|
|
if (dev->test_device)
|
|
|
|
|
{
|
|
|
|
|
struct ether_addr good_addr = { {0x70, 0x37, 0x03, 0x70, 0x37, 0x03} };
|
|
|
|
|
struct ether_addr bad_addr = { {0x00, 0x00, 0x00, 0x00, 0x00, 0x00} };
|
|
|
|
|
gboolean link = nm_device_get_link_active (dev);
|
|
|
|
|
|
|
|
|
|
memcpy ((link ? &good_addr : &bad_addr), &(wrq.u.ap_addr.sa_data), sizeof (struct ether_addr));
|
|
|
|
|
return;
|
|
|
|
|
}
|
|
|
|
|
|
2004-07-06 01:34:10 +00:00
|
|
|
iwlib_socket = iw_sockets_open ();
|
|
|
|
|
if (iw_get_ext (iwlib_socket, nm_device_get_iface (dev), SIOCGIWAP, &wrq) >= 0)
|
|
|
|
|
memcpy (addr, &(wrq.u.ap_addr.sa_data), sizeof (struct ether_addr));
|
|
|
|
|
else
|
|
|
|
|
memset (addr, 0, sizeof (struct ether_addr));
|
|
|
|
|
close (iwlib_socket);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
2004-06-24 14:18:37 +00:00
|
|
|
/*
|
2004-08-24 Dan Williams <dcbw@redhat.com>
* src/NetworkManagerAP.[ch]
- Add a "enc_method_good" member and accessors to an Access Point
to signal when we've found the correct encryption method
for an access point
- Add a "timestamp" member and accessors, remove "priority" member
and accessors (use timestamps instead)
- Rename "wep_key"->"enc_key"
- (nm_ap_get_enc_key_hashed): new, return the correct mangled key
for a specified encryption method using the access points
source encryption key/passphrase
* src/NetworkManagerAPList.c
- When updating a network with dbus, grab timestamp now instead of
priority
* src/NetworkManagerDBus.[ch]
- Add signal for "DeviceActivating"
- Switch priority->timestamp
* src/NetworkManagerDevice.c
- Change references of "wep_key" -> "enc_key" or "key"
- Signal DeviceActivating when starting activation
- When activating a wireless device, if the access point we are connecting
to is encrypted, and we have a source key, try to generate a mangled
key and use that (ie, generate real WEP key from a passphrase)
- Rework device activation to fallback to other encryption methods if
a previous one didn't work (ie, try mangling a key as a 104-bit passphrase
first, then if that doesn't work fall back to direct hex key).
- (nm_device_update_best_ap): fix a deadlock, and use timestamps instead of
priority. We now prefer the latest access point used, rather than using
a priority scheme
- (nm_device_do_normal_scan): make the encryption method "unknown" on access
points we've just discovered, and merge in correct info from the global
access point lists
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@68 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-25 22:41:12 +00:00
|
|
|
* nm_device_set_enc_key
|
2004-06-24 14:18:37 +00:00
|
|
|
*
|
2004-08-24 Dan Williams <dcbw@redhat.com>
* src/NetworkManagerAP.[ch]
- Add a "enc_method_good" member and accessors to an Access Point
to signal when we've found the correct encryption method
for an access point
- Add a "timestamp" member and accessors, remove "priority" member
and accessors (use timestamps instead)
- Rename "wep_key"->"enc_key"
- (nm_ap_get_enc_key_hashed): new, return the correct mangled key
for a specified encryption method using the access points
source encryption key/passphrase
* src/NetworkManagerAPList.c
- When updating a network with dbus, grab timestamp now instead of
priority
* src/NetworkManagerDBus.[ch]
- Add signal for "DeviceActivating"
- Switch priority->timestamp
* src/NetworkManagerDevice.c
- Change references of "wep_key" -> "enc_key" or "key"
- Signal DeviceActivating when starting activation
- When activating a wireless device, if the access point we are connecting
to is encrypted, and we have a source key, try to generate a mangled
key and use that (ie, generate real WEP key from a passphrase)
- Rework device activation to fallback to other encryption methods if
a previous one didn't work (ie, try mangling a key as a 104-bit passphrase
first, then if that doesn't work fall back to direct hex key).
- (nm_device_update_best_ap): fix a deadlock, and use timestamps instead of
priority. We now prefer the latest access point used, rather than using
a priority scheme
- (nm_device_do_normal_scan): make the encryption method "unknown" on access
points we've just discovered, and merge in correct info from the global
access point lists
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@68 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-25 22:41:12 +00:00
|
|
|
* If a device is wireless, set the encryption key that it should use.
|
2004-06-24 14:18:37 +00:00
|
|
|
*
|
2004-08-24 Dan Williams <dcbw@redhat.com>
* src/NetworkManagerAP.[ch]
- Add a "enc_method_good" member and accessors to an Access Point
to signal when we've found the correct encryption method
for an access point
- Add a "timestamp" member and accessors, remove "priority" member
and accessors (use timestamps instead)
- Rename "wep_key"->"enc_key"
- (nm_ap_get_enc_key_hashed): new, return the correct mangled key
for a specified encryption method using the access points
source encryption key/passphrase
* src/NetworkManagerAPList.c
- When updating a network with dbus, grab timestamp now instead of
priority
* src/NetworkManagerDBus.[ch]
- Add signal for "DeviceActivating"
- Switch priority->timestamp
* src/NetworkManagerDevice.c
- Change references of "wep_key" -> "enc_key" or "key"
- Signal DeviceActivating when starting activation
- When activating a wireless device, if the access point we are connecting
to is encrypted, and we have a source key, try to generate a mangled
key and use that (ie, generate real WEP key from a passphrase)
- Rework device activation to fallback to other encryption methods if
a previous one didn't work (ie, try mangling a key as a 104-bit passphrase
first, then if that doesn't work fall back to direct hex key).
- (nm_device_update_best_ap): fix a deadlock, and use timestamps instead of
priority. We now prefer the latest access point used, rather than using
a priority scheme
- (nm_device_do_normal_scan): make the encryption method "unknown" on access
points we've just discovered, and merge in correct info from the global
access point lists
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@68 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-25 22:41:12 +00:00
|
|
|
* key: encryption key to use, or NULL or "" to disable encryption.
|
|
|
|
|
* NOTE that at this time, the key must be the raw HEX key, not
|
|
|
|
|
* a passphrase.
|
2004-06-24 14:18:37 +00:00
|
|
|
*/
|
2004-11-03 07:47:32 +00:00
|
|
|
void nm_device_set_enc_key (NMDevice *dev, const char *key, NMDeviceAuthMethod auth_method)
|
2004-06-24 14:18:37 +00:00
|
|
|
{
|
|
|
|
|
int iwlib_socket;
|
|
|
|
|
int err;
|
|
|
|
|
struct iwreq wreq;
|
|
|
|
|
int keylen;
|
2004-07-06 04:45:00 +00:00
|
|
|
unsigned char safe_key[IW_ENCODING_TOKEN_MAX + 1];
|
2004-06-24 14:18:37 +00:00
|
|
|
gboolean set_key = FALSE;
|
|
|
|
|
|
|
|
|
|
g_return_if_fail (dev != NULL);
|
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
|
|
|
g_return_if_fail (nm_device_is_wireless (dev));
|
2004-06-24 14:18:37 +00:00
|
|
|
|
2004-08-31 Dan Williams <dcbw@redhat.com>
* Remove 'debug' extern global from all files since we now
use syslog()
* src/NetworkManager.[ch]
- Break out routine that get the net.interface property from HAL,
removing that logic from nm_create_device_and_add_to_list()
- (nm_create_device_and_add_to_list): make this a bit more general so
it doesn't do the talking to HAL. Also add arguments to facilitate
the create of test devices.
- (nm_data_mark_state_changed): rename from nm_data_set_state_modified()
- (nm_data_new, main, nm_print_usage): add new argument "--enable-test-devices"
which makes NetworkManager listen for dbus commands to create test
devices, which have no backing hardware. Use when you're on a plane
for example, and/or forgot your wireless card at home. Test devices
_cannot_ be created unless NM is started with --enable-test-devices.
* src/NetworkManagerDbus.[ch]
- New "getLinkActive" method for devices
- New "setLinkActive" method for devices (only works on test devices)
- New "createTestDevice" method on NetworkManager object to create a test
device of a specified type (ie wired, wireless). UDI is created from
scratch, as is the interface name. Only works when NM is started with
--enable-test-devices switch.
- New "removeTestDevice" method on NetworkManager object which removes a
test device. Only works when NM is started with --enable-test-devices
* src/NetworkManagerDevice.[ch]
- Logic to facilitate test devices. Add variables to NMDevice struct to indicate
whether a device is a test device or not, and what its link status is.
- Deal with test devices in most functions. For those that work directly on hardware
special-case test devices.
- (nm_device_new): don't create a test device if test devices weren't enabled on the
command-line.
- (nm_device_update_link_active): split out logic for wired and wireless device link
checking to separate functions to facilitate test device link checking.
- (nm_device_set_enc_key): Since some drivers for wireless cards are daft and
don't make a distinction between System Authentication and Encryption
(namely Cisco aironet), we use Open System auth when setting a WEP key
on the card. We don't deal with Shared Key auth yet.
- (nm_device_activation_worker): split the activation cancel check logic out into
a separate routine nm_device_activation_cancel_if_needed()
- (nm_device_activation_signal_cancel): rename from nm_device_activation_cancel()
- (nm_device_fake_ap_list): Test wireless devices obviously cannot scan, so create
a list of fake access points that they can "see"
- (nm_device_is_test_device): return whether or not a device is a test device
* src/NetworkManagerPolicy.c
- (nm_policy_get_best_device): attempt to deal with wireless network selection,
previously if you "locked"/forced NM to use a wireless device but then
selected a wireless network for NM to use, it would switch to a wired device.
So, if the active device is wireless and it has a "forced" best AP, use it
if the "forced" best AP is still valid
- (nm_state_modification_monitor): deal with NULL best devices, for example
there were no usable network devices, or the last one was removed
* src/backends/NetworkManager*.c
- Deal with test devices, mostly just return success for operations like getting
a DHCP address
* test/nmtestdevices.c
- Test tool to create/remove/link-switch test devices
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@112 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-31 16:09:15 +00:00
|
|
|
/* Test devices just ignore encryption keys */
|
|
|
|
|
if (dev->test_device)
|
|
|
|
|
return;
|
|
|
|
|
|
2004-06-24 14:18:37 +00:00
|
|
|
/* Make sure the essid we get passed is a valid size */
|
2004-08-24 Dan Williams <dcbw@redhat.com>
* src/NetworkManagerAP.[ch]
- Add a "enc_method_good" member and accessors to an Access Point
to signal when we've found the correct encryption method
for an access point
- Add a "timestamp" member and accessors, remove "priority" member
and accessors (use timestamps instead)
- Rename "wep_key"->"enc_key"
- (nm_ap_get_enc_key_hashed): new, return the correct mangled key
for a specified encryption method using the access points
source encryption key/passphrase
* src/NetworkManagerAPList.c
- When updating a network with dbus, grab timestamp now instead of
priority
* src/NetworkManagerDBus.[ch]
- Add signal for "DeviceActivating"
- Switch priority->timestamp
* src/NetworkManagerDevice.c
- Change references of "wep_key" -> "enc_key" or "key"
- Signal DeviceActivating when starting activation
- When activating a wireless device, if the access point we are connecting
to is encrypted, and we have a source key, try to generate a mangled
key and use that (ie, generate real WEP key from a passphrase)
- Rework device activation to fallback to other encryption methods if
a previous one didn't work (ie, try mangling a key as a 104-bit passphrase
first, then if that doesn't work fall back to direct hex key).
- (nm_device_update_best_ap): fix a deadlock, and use timestamps instead of
priority. We now prefer the latest access point used, rather than using
a priority scheme
- (nm_device_do_normal_scan): make the encryption method "unknown" on access
points we've just discovered, and merge in correct info from the global
access point lists
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@68 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-25 22:41:12 +00:00
|
|
|
if (!key)
|
2004-06-24 14:18:37 +00:00
|
|
|
safe_key[0] = '\0';
|
|
|
|
|
else
|
|
|
|
|
{
|
2004-08-24 Dan Williams <dcbw@redhat.com>
* src/NetworkManagerAP.[ch]
- Add a "enc_method_good" member and accessors to an Access Point
to signal when we've found the correct encryption method
for an access point
- Add a "timestamp" member and accessors, remove "priority" member
and accessors (use timestamps instead)
- Rename "wep_key"->"enc_key"
- (nm_ap_get_enc_key_hashed): new, return the correct mangled key
for a specified encryption method using the access points
source encryption key/passphrase
* src/NetworkManagerAPList.c
- When updating a network with dbus, grab timestamp now instead of
priority
* src/NetworkManagerDBus.[ch]
- Add signal for "DeviceActivating"
- Switch priority->timestamp
* src/NetworkManagerDevice.c
- Change references of "wep_key" -> "enc_key" or "key"
- Signal DeviceActivating when starting activation
- When activating a wireless device, if the access point we are connecting
to is encrypted, and we have a source key, try to generate a mangled
key and use that (ie, generate real WEP key from a passphrase)
- Rework device activation to fallback to other encryption methods if
a previous one didn't work (ie, try mangling a key as a 104-bit passphrase
first, then if that doesn't work fall back to direct hex key).
- (nm_device_update_best_ap): fix a deadlock, and use timestamps instead of
priority. We now prefer the latest access point used, rather than using
a priority scheme
- (nm_device_do_normal_scan): make the encryption method "unknown" on access
points we've just discovered, and merge in correct info from the global
access point lists
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@68 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-25 22:41:12 +00:00
|
|
|
strncpy (safe_key, key, IW_ENCODING_TOKEN_MAX);
|
2004-06-24 14:18:37 +00:00
|
|
|
safe_key[IW_ENCODING_TOKEN_MAX] = '\0';
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
iwlib_socket = iw_sockets_open ();
|
|
|
|
|
if (iwlib_socket >= 0)
|
|
|
|
|
{
|
|
|
|
|
wreq.u.data.pointer = (caddr_t) NULL;
|
|
|
|
|
wreq.u.data.length = 0;
|
2004-08-31 Dan Williams <dcbw@redhat.com>
* Remove 'debug' extern global from all files since we now
use syslog()
* src/NetworkManager.[ch]
- Break out routine that get the net.interface property from HAL,
removing that logic from nm_create_device_and_add_to_list()
- (nm_create_device_and_add_to_list): make this a bit more general so
it doesn't do the talking to HAL. Also add arguments to facilitate
the create of test devices.
- (nm_data_mark_state_changed): rename from nm_data_set_state_modified()
- (nm_data_new, main, nm_print_usage): add new argument "--enable-test-devices"
which makes NetworkManager listen for dbus commands to create test
devices, which have no backing hardware. Use when you're on a plane
for example, and/or forgot your wireless card at home. Test devices
_cannot_ be created unless NM is started with --enable-test-devices.
* src/NetworkManagerDbus.[ch]
- New "getLinkActive" method for devices
- New "setLinkActive" method for devices (only works on test devices)
- New "createTestDevice" method on NetworkManager object to create a test
device of a specified type (ie wired, wireless). UDI is created from
scratch, as is the interface name. Only works when NM is started with
--enable-test-devices switch.
- New "removeTestDevice" method on NetworkManager object which removes a
test device. Only works when NM is started with --enable-test-devices
* src/NetworkManagerDevice.[ch]
- Logic to facilitate test devices. Add variables to NMDevice struct to indicate
whether a device is a test device or not, and what its link status is.
- Deal with test devices in most functions. For those that work directly on hardware
special-case test devices.
- (nm_device_new): don't create a test device if test devices weren't enabled on the
command-line.
- (nm_device_update_link_active): split out logic for wired and wireless device link
checking to separate functions to facilitate test device link checking.
- (nm_device_set_enc_key): Since some drivers for wireless cards are daft and
don't make a distinction between System Authentication and Encryption
(namely Cisco aironet), we use Open System auth when setting a WEP key
on the card. We don't deal with Shared Key auth yet.
- (nm_device_activation_worker): split the activation cancel check logic out into
a separate routine nm_device_activation_cancel_if_needed()
- (nm_device_activation_signal_cancel): rename from nm_device_activation_cancel()
- (nm_device_fake_ap_list): Test wireless devices obviously cannot scan, so create
a list of fake access points that they can "see"
- (nm_device_is_test_device): return whether or not a device is a test device
* src/NetworkManagerPolicy.c
- (nm_policy_get_best_device): attempt to deal with wireless network selection,
previously if you "locked"/forced NM to use a wireless device but then
selected a wireless network for NM to use, it would switch to a wired device.
So, if the active device is wireless and it has a "forced" best AP, use it
if the "forced" best AP is still valid
- (nm_state_modification_monitor): deal with NULL best devices, for example
there were no usable network devices, or the last one was removed
* src/backends/NetworkManager*.c
- Deal with test devices, mostly just return success for operations like getting
a DHCP address
* test/nmtestdevices.c
- Test tool to create/remove/link-switch test devices
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@112 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-31 16:09:15 +00:00
|
|
|
wreq.u.data.flags = IW_ENCODE_ENABLED;
|
|
|
|
|
|
|
|
|
|
/* Unfortunately, some drivers (Cisco) don't make a distinction between
|
|
|
|
|
* Open System authentication mode and whether or not to use WEP. You
|
|
|
|
|
* DON'T have to use WEP when using Open System, but these cards force
|
|
|
|
|
* it. Therefore, we have to set Open System mode when using WEP.
|
|
|
|
|
*/
|
2004-06-24 14:18:37 +00:00
|
|
|
|
|
|
|
|
if (strlen (safe_key) == 0)
|
|
|
|
|
{
|
2004-08-31 Dan Williams <dcbw@redhat.com>
* Remove 'debug' extern global from all files since we now
use syslog()
* src/NetworkManager.[ch]
- Break out routine that get the net.interface property from HAL,
removing that logic from nm_create_device_and_add_to_list()
- (nm_create_device_and_add_to_list): make this a bit more general so
it doesn't do the talking to HAL. Also add arguments to facilitate
the create of test devices.
- (nm_data_mark_state_changed): rename from nm_data_set_state_modified()
- (nm_data_new, main, nm_print_usage): add new argument "--enable-test-devices"
which makes NetworkManager listen for dbus commands to create test
devices, which have no backing hardware. Use when you're on a plane
for example, and/or forgot your wireless card at home. Test devices
_cannot_ be created unless NM is started with --enable-test-devices.
* src/NetworkManagerDbus.[ch]
- New "getLinkActive" method for devices
- New "setLinkActive" method for devices (only works on test devices)
- New "createTestDevice" method on NetworkManager object to create a test
device of a specified type (ie wired, wireless). UDI is created from
scratch, as is the interface name. Only works when NM is started with
--enable-test-devices switch.
- New "removeTestDevice" method on NetworkManager object which removes a
test device. Only works when NM is started with --enable-test-devices
* src/NetworkManagerDevice.[ch]
- Logic to facilitate test devices. Add variables to NMDevice struct to indicate
whether a device is a test device or not, and what its link status is.
- Deal with test devices in most functions. For those that work directly on hardware
special-case test devices.
- (nm_device_new): don't create a test device if test devices weren't enabled on the
command-line.
- (nm_device_update_link_active): split out logic for wired and wireless device link
checking to separate functions to facilitate test device link checking.
- (nm_device_set_enc_key): Since some drivers for wireless cards are daft and
don't make a distinction between System Authentication and Encryption
(namely Cisco aironet), we use Open System auth when setting a WEP key
on the card. We don't deal with Shared Key auth yet.
- (nm_device_activation_worker): split the activation cancel check logic out into
a separate routine nm_device_activation_cancel_if_needed()
- (nm_device_activation_signal_cancel): rename from nm_device_activation_cancel()
- (nm_device_fake_ap_list): Test wireless devices obviously cannot scan, so create
a list of fake access points that they can "see"
- (nm_device_is_test_device): return whether or not a device is a test device
* src/NetworkManagerPolicy.c
- (nm_policy_get_best_device): attempt to deal with wireless network selection,
previously if you "locked"/forced NM to use a wireless device but then
selected a wireless network for NM to use, it would switch to a wired device.
So, if the active device is wireless and it has a "forced" best AP, use it
if the "forced" best AP is still valid
- (nm_state_modification_monitor): deal with NULL best devices, for example
there were no usable network devices, or the last one was removed
* src/backends/NetworkManager*.c
- Deal with test devices, mostly just return success for operations like getting
a DHCP address
* test/nmtestdevices.c
- Test tool to create/remove/link-switch test devices
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@112 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-31 16:09:15 +00:00
|
|
|
wreq.u.data.flags |= IW_ENCODE_DISABLED | IW_ENCODE_NOKEY;
|
2004-06-24 14:18:37 +00:00
|
|
|
set_key = TRUE;
|
|
|
|
|
}
|
|
|
|
|
else
|
|
|
|
|
{
|
2004-07-06 04:45:00 +00:00
|
|
|
unsigned char parsed_key[IW_ENCODING_TOKEN_MAX + 1];
|
|
|
|
|
|
|
|
|
|
keylen = iw_in_key_full(iwlib_socket, nm_device_get_iface (dev), safe_key, &parsed_key[0], &wreq.u.data.flags);
|
2004-06-24 14:18:37 +00:00
|
|
|
if (keylen > 0)
|
|
|
|
|
{
|
2004-11-03 07:47:32 +00:00
|
|
|
switch (auth_method)
|
|
|
|
|
{
|
|
|
|
|
case NM_DEVICE_AUTH_METHOD_OPEN_SYSTEM:
|
|
|
|
|
wreq.u.data.flags |= IW_ENCODE_OPEN;
|
|
|
|
|
break;
|
|
|
|
|
case NM_DEVICE_AUTH_METHOD_SHARED_KEY:
|
|
|
|
|
wreq.u.data.flags |= IW_ENCODE_RESTRICTED;
|
|
|
|
|
break;
|
|
|
|
|
default:
|
|
|
|
|
wreq.u.data.flags |= IW_ENCODE_RESTRICTED;
|
|
|
|
|
break;
|
|
|
|
|
}
|
2004-07-06 04:45:00 +00:00
|
|
|
wreq.u.data.pointer = (caddr_t) &parsed_key;
|
2004-06-24 14:18:37 +00:00
|
|
|
wreq.u.data.length = keylen;
|
|
|
|
|
set_key = TRUE;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (set_key)
|
|
|
|
|
{
|
|
|
|
|
err = iw_set_ext (iwlib_socket, nm_device_get_iface (dev), SIOCSIWENCODE, &wreq);
|
|
|
|
|
if (err == -1)
|
2004-08-24 Dan Williams <dcbw@redhat.com>
* src/NetworkManagerAP.[ch]
- Add a "enc_method_good" member and accessors to an Access Point
to signal when we've found the correct encryption method
for an access point
- Add a "timestamp" member and accessors, remove "priority" member
and accessors (use timestamps instead)
- Rename "wep_key"->"enc_key"
- (nm_ap_get_enc_key_hashed): new, return the correct mangled key
for a specified encryption method using the access points
source encryption key/passphrase
* src/NetworkManagerAPList.c
- When updating a network with dbus, grab timestamp now instead of
priority
* src/NetworkManagerDBus.[ch]
- Add signal for "DeviceActivating"
- Switch priority->timestamp
* src/NetworkManagerDevice.c
- Change references of "wep_key" -> "enc_key" or "key"
- Signal DeviceActivating when starting activation
- When activating a wireless device, if the access point we are connecting
to is encrypted, and we have a source key, try to generate a mangled
key and use that (ie, generate real WEP key from a passphrase)
- Rework device activation to fallback to other encryption methods if
a previous one didn't work (ie, try mangling a key as a 104-bit passphrase
first, then if that doesn't work fall back to direct hex key).
- (nm_device_update_best_ap): fix a deadlock, and use timestamps instead of
priority. We now prefer the latest access point used, rather than using
a priority scheme
- (nm_device_do_normal_scan): make the encryption method "unknown" on access
points we've just discovered, and merge in correct info from the global
access point lists
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@68 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-25 22:41:12 +00:00
|
|
|
syslog (LOG_ERR, "nm_device_set_enc_key(): error setting key for device %s. errno = %d", nm_device_get_iface (dev), errno);
|
2004-06-24 14:18:37 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
close (iwlib_socket);
|
2004-08-24 Dan Williams <dcbw@redhat.com>
* src/NetworkManagerAP.[ch]
- Add a "enc_method_good" member and accessors to an Access Point
to signal when we've found the correct encryption method
for an access point
- Add a "timestamp" member and accessors, remove "priority" member
and accessors (use timestamps instead)
- Rename "wep_key"->"enc_key"
- (nm_ap_get_enc_key_hashed): new, return the correct mangled key
for a specified encryption method using the access points
source encryption key/passphrase
* src/NetworkManagerAPList.c
- When updating a network with dbus, grab timestamp now instead of
priority
* src/NetworkManagerDBus.[ch]
- Add signal for "DeviceActivating"
- Switch priority->timestamp
* src/NetworkManagerDevice.c
- Change references of "wep_key" -> "enc_key" or "key"
- Signal DeviceActivating when starting activation
- When activating a wireless device, if the access point we are connecting
to is encrypted, and we have a source key, try to generate a mangled
key and use that (ie, generate real WEP key from a passphrase)
- Rework device activation to fallback to other encryption methods if
a previous one didn't work (ie, try mangling a key as a 104-bit passphrase
first, then if that doesn't work fall back to direct hex key).
- (nm_device_update_best_ap): fix a deadlock, and use timestamps instead of
priority. We now prefer the latest access point used, rather than using
a priority scheme
- (nm_device_do_normal_scan): make the encryption method "unknown" on access
points we've just discovered, and merge in correct info from the global
access point lists
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@68 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-25 22:41:12 +00:00
|
|
|
} else syslog (LOG_ERR, "nm_device_set_enc_key(): could not get wireless control socket.");
|
2004-06-24 14:18:37 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
2004-09-13 17:43:16 +00:00
|
|
|
/*
|
|
|
|
|
* nm_device_get_signal_strength
|
|
|
|
|
*
|
|
|
|
|
* Get the current signal strength of a wireless device. This only works when
|
|
|
|
|
* the card is associated with an access point, so will only work for the
|
|
|
|
|
* active device.
|
|
|
|
|
*
|
|
|
|
|
* Returns: -1 on error
|
|
|
|
|
* 0 - 100 strength percentage of the connection to the current access point
|
|
|
|
|
*
|
|
|
|
|
*/
|
|
|
|
|
gint8 nm_device_get_signal_strength (NMDevice *dev)
|
|
|
|
|
{
|
|
|
|
|
g_return_val_if_fail (dev != NULL, -1);
|
|
|
|
|
g_return_val_if_fail (nm_device_is_wireless (dev), -1);
|
|
|
|
|
|
|
|
|
|
return (dev->options.wireless.strength);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* nm_device_update_signal_strength
|
|
|
|
|
*
|
|
|
|
|
* Update the device's idea of the strength of its connection to the
|
|
|
|
|
* current access point.
|
|
|
|
|
*
|
|
|
|
|
*/
|
|
|
|
|
void nm_device_update_signal_strength (NMDevice *dev)
|
|
|
|
|
{
|
|
|
|
|
gboolean has_range;
|
|
|
|
|
int iwlib_socket;
|
|
|
|
|
iwrange range;
|
|
|
|
|
iwstats stats;
|
|
|
|
|
int percent = -1;
|
|
|
|
|
|
|
|
|
|
g_return_if_fail (dev != NULL);
|
|
|
|
|
g_return_if_fail (nm_device_is_wireless (dev));
|
|
|
|
|
g_return_if_fail (dev->app_data != NULL);
|
|
|
|
|
|
|
|
|
|
/* If we aren't the active device, we don't really have a signal strength
|
|
|
|
|
* that would mean anything.
|
|
|
|
|
*/
|
|
|
|
|
#if 0
|
|
|
|
|
if (dev != dev->app_data->active_device)
|
|
|
|
|
{
|
|
|
|
|
dev->options.wireless.strength = -1;
|
|
|
|
|
return;
|
|
|
|
|
}
|
|
|
|
|
#endif
|
|
|
|
|
|
|
|
|
|
/* Fake a value for test devices */
|
|
|
|
|
if (dev->test_device)
|
|
|
|
|
{
|
|
|
|
|
dev->options.wireless.strength = 75;
|
|
|
|
|
return;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
iwlib_socket = iw_sockets_open ();
|
|
|
|
|
has_range = (iw_get_range_info (iwlib_socket, nm_device_get_iface (dev), &range) >= 0);
|
|
|
|
|
if (iw_get_stats (iwlib_socket, nm_device_get_iface (dev), &stats, &range, has_range) == 0)
|
|
|
|
|
{
|
|
|
|
|
/* Update our max quality while we're at it */
|
|
|
|
|
dev->options.wireless.max_quality = range.max_qual.level;
|
|
|
|
|
dev->options.wireless.noise = stats.qual.noise;
|
|
|
|
|
percent = nm_wireless_qual_to_percent (dev, &(stats.qual));
|
|
|
|
|
}
|
|
|
|
|
else
|
|
|
|
|
{
|
|
|
|
|
dev->options.wireless.max_quality = -1;
|
|
|
|
|
dev->options.wireless.noise = -1;
|
|
|
|
|
percent = -1;
|
|
|
|
|
}
|
|
|
|
|
close (iwlib_socket);
|
|
|
|
|
|
|
|
|
|
dev->options.wireless.strength = percent;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* nm_device_get_noise
|
|
|
|
|
*
|
|
|
|
|
* Get the current noise level of a wireless device.
|
|
|
|
|
*
|
|
|
|
|
*/
|
|
|
|
|
guint8 nm_device_get_noise (NMDevice *dev)
|
|
|
|
|
{
|
|
|
|
|
g_return_val_if_fail (dev != NULL, 0);
|
|
|
|
|
g_return_val_if_fail (nm_device_is_wireless (dev), 0);
|
|
|
|
|
|
|
|
|
|
return (dev->options.wireless.noise);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
2004-08-26 20:05:24 +00:00
|
|
|
/*
|
|
|
|
|
* nm_device_get_max_quality
|
|
|
|
|
*
|
2004-09-13 17:43:16 +00:00
|
|
|
* Get the quality maximum of a wireless device.
|
2004-08-26 20:05:24 +00:00
|
|
|
*
|
|
|
|
|
*/
|
|
|
|
|
guint8 nm_device_get_max_quality (NMDevice *dev)
|
|
|
|
|
{
|
|
|
|
|
g_return_val_if_fail (dev != NULL, 0);
|
|
|
|
|
g_return_val_if_fail (nm_device_is_wireless (dev), 0);
|
|
|
|
|
|
|
|
|
|
return (dev->options.wireless.max_quality);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
2004-10-04 15:44:00 +00:00
|
|
|
/*
|
|
|
|
|
* nm_device_get_bad_crypt_packets
|
|
|
|
|
*
|
|
|
|
|
* Return the number of packets the card has dropped because
|
|
|
|
|
* they could not be successfully decrypted.
|
|
|
|
|
*
|
|
|
|
|
*/
|
|
|
|
|
guint32 nm_device_get_bad_crypt_packets (NMDevice *dev)
|
|
|
|
|
{
|
|
|
|
|
iwstats stats;
|
|
|
|
|
int sk;
|
|
|
|
|
int err;
|
|
|
|
|
|
|
|
|
|
g_return_val_if_fail (dev != NULL, 0);
|
|
|
|
|
g_return_val_if_fail (nm_device_is_wireless (dev), 0);
|
|
|
|
|
|
|
|
|
|
sk = iw_sockets_open ();
|
|
|
|
|
err = iw_get_stats (sk, nm_device_get_iface (dev), &stats, NULL, FALSE);
|
|
|
|
|
close (sk);
|
|
|
|
|
return (err == 0 ? stats.discard.code : 0);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
2004-07-06 01:34:10 +00:00
|
|
|
/*
|
|
|
|
|
* nm_device_get_ip4_address
|
|
|
|
|
*
|
|
|
|
|
* Get a device's IPv4 address
|
|
|
|
|
*
|
|
|
|
|
*/
|
|
|
|
|
guint32 nm_device_get_ip4_address(NMDevice *dev)
|
|
|
|
|
{
|
|
|
|
|
g_return_val_if_fail (dev != NULL, 0);
|
2004-07-06 04:45:00 +00:00
|
|
|
|
|
|
|
|
return (dev->ip4_address);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
void nm_device_update_ip4_address (NMDevice *dev)
|
|
|
|
|
{
|
2004-07-27 16:15:36 +00:00
|
|
|
guint32 new_address;
|
|
|
|
|
struct ifreq req;
|
|
|
|
|
int socket;
|
|
|
|
|
int err;
|
2004-07-06 04:45:00 +00:00
|
|
|
|
|
|
|
|
g_return_if_fail (dev != NULL);
|
2004-07-25 02:40:19 +00:00
|
|
|
g_return_if_fail (dev->app_data != NULL);
|
2004-07-06 04:45:00 +00:00
|
|
|
g_return_if_fail (nm_device_get_iface (dev) != NULL);
|
2004-07-06 01:34:10 +00:00
|
|
|
|
2004-08-31 Dan Williams <dcbw@redhat.com>
* Remove 'debug' extern global from all files since we now
use syslog()
* src/NetworkManager.[ch]
- Break out routine that get the net.interface property from HAL,
removing that logic from nm_create_device_and_add_to_list()
- (nm_create_device_and_add_to_list): make this a bit more general so
it doesn't do the talking to HAL. Also add arguments to facilitate
the create of test devices.
- (nm_data_mark_state_changed): rename from nm_data_set_state_modified()
- (nm_data_new, main, nm_print_usage): add new argument "--enable-test-devices"
which makes NetworkManager listen for dbus commands to create test
devices, which have no backing hardware. Use when you're on a plane
for example, and/or forgot your wireless card at home. Test devices
_cannot_ be created unless NM is started with --enable-test-devices.
* src/NetworkManagerDbus.[ch]
- New "getLinkActive" method for devices
- New "setLinkActive" method for devices (only works on test devices)
- New "createTestDevice" method on NetworkManager object to create a test
device of a specified type (ie wired, wireless). UDI is created from
scratch, as is the interface name. Only works when NM is started with
--enable-test-devices switch.
- New "removeTestDevice" method on NetworkManager object which removes a
test device. Only works when NM is started with --enable-test-devices
* src/NetworkManagerDevice.[ch]
- Logic to facilitate test devices. Add variables to NMDevice struct to indicate
whether a device is a test device or not, and what its link status is.
- Deal with test devices in most functions. For those that work directly on hardware
special-case test devices.
- (nm_device_new): don't create a test device if test devices weren't enabled on the
command-line.
- (nm_device_update_link_active): split out logic for wired and wireless device link
checking to separate functions to facilitate test device link checking.
- (nm_device_set_enc_key): Since some drivers for wireless cards are daft and
don't make a distinction between System Authentication and Encryption
(namely Cisco aironet), we use Open System auth when setting a WEP key
on the card. We don't deal with Shared Key auth yet.
- (nm_device_activation_worker): split the activation cancel check logic out into
a separate routine nm_device_activation_cancel_if_needed()
- (nm_device_activation_signal_cancel): rename from nm_device_activation_cancel()
- (nm_device_fake_ap_list): Test wireless devices obviously cannot scan, so create
a list of fake access points that they can "see"
- (nm_device_is_test_device): return whether or not a device is a test device
* src/NetworkManagerPolicy.c
- (nm_policy_get_best_device): attempt to deal with wireless network selection,
previously if you "locked"/forced NM to use a wireless device but then
selected a wireless network for NM to use, it would switch to a wired device.
So, if the active device is wireless and it has a "forced" best AP, use it
if the "forced" best AP is still valid
- (nm_state_modification_monitor): deal with NULL best devices, for example
there were no usable network devices, or the last one was removed
* src/backends/NetworkManager*.c
- Deal with test devices, mostly just return success for operations like getting
a DHCP address
* test/nmtestdevices.c
- Test tool to create/remove/link-switch test devices
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@112 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-31 16:09:15 +00:00
|
|
|
/* Test devices get a nice, bogus IP address */
|
|
|
|
|
if (dev->test_device)
|
|
|
|
|
{
|
|
|
|
|
dev->ip4_address = 0x07030703;
|
|
|
|
|
return;
|
|
|
|
|
}
|
|
|
|
|
|
2004-07-06 01:34:10 +00:00
|
|
|
socket = nm_get_network_control_socket ();
|
|
|
|
|
if (socket < 0)
|
2004-07-06 04:45:00 +00:00
|
|
|
return;
|
2004-07-06 01:34:10 +00:00
|
|
|
|
2004-10-23 06:43:01 +00:00
|
|
|
strncpy ((char *)(&req.ifr_name), nm_device_get_iface (dev), 16); /* 16 == IF_NAMESIZE */
|
2004-07-27 16:15:36 +00:00
|
|
|
err = ioctl (socket, SIOCGIFADDR, &req);
|
|
|
|
|
close (socket);
|
|
|
|
|
if (err != 0)
|
2004-07-06 04:45:00 +00:00
|
|
|
return;
|
2004-07-06 01:34:10 +00:00
|
|
|
|
2004-07-06 04:45:00 +00:00
|
|
|
new_address = ((struct sockaddr_in *)(&req.ifr_addr))->sin_addr.s_addr;
|
|
|
|
|
|
|
|
|
|
/* If the new address is different, send an IP4AddressChanged signal on the bus */
|
|
|
|
|
if (new_address != nm_device_get_ip4_address (dev))
|
|
|
|
|
{
|
2004-07-25 02:40:19 +00:00
|
|
|
nm_dbus_signal_device_ip4_address_change (dev->app_data->dbus_connection, dev);
|
2004-07-06 04:45:00 +00:00
|
|
|
dev->ip4_address = new_address;
|
|
|
|
|
}
|
2004-07-06 01:34:10 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* nm_device_get_ip6_address
|
|
|
|
|
*
|
|
|
|
|
* Get a device's IPv6 address
|
|
|
|
|
*
|
|
|
|
|
*/
|
|
|
|
|
void nm_device_get_ip6_address(NMDevice *dev)
|
|
|
|
|
{
|
|
|
|
|
/* FIXME
|
|
|
|
|
* Implement
|
|
|
|
|
*/
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
2004-06-24 14:18:37 +00:00
|
|
|
/*
|
|
|
|
|
* nm_device_set_up_down
|
|
|
|
|
*
|
|
|
|
|
* Set the up flag on the device on or off
|
|
|
|
|
*
|
|
|
|
|
*/
|
|
|
|
|
static void nm_device_set_up_down (NMDevice *dev, gboolean up)
|
|
|
|
|
{
|
|
|
|
|
struct ifreq ifr;
|
|
|
|
|
int iface_fd;
|
|
|
|
|
int err;
|
|
|
|
|
guint32 flags = up ? IFF_UP : ~IFF_UP;
|
|
|
|
|
|
|
|
|
|
g_return_if_fail (dev != NULL);
|
|
|
|
|
|
2004-08-31 Dan Williams <dcbw@redhat.com>
* Remove 'debug' extern global from all files since we now
use syslog()
* src/NetworkManager.[ch]
- Break out routine that get the net.interface property from HAL,
removing that logic from nm_create_device_and_add_to_list()
- (nm_create_device_and_add_to_list): make this a bit more general so
it doesn't do the talking to HAL. Also add arguments to facilitate
the create of test devices.
- (nm_data_mark_state_changed): rename from nm_data_set_state_modified()
- (nm_data_new, main, nm_print_usage): add new argument "--enable-test-devices"
which makes NetworkManager listen for dbus commands to create test
devices, which have no backing hardware. Use when you're on a plane
for example, and/or forgot your wireless card at home. Test devices
_cannot_ be created unless NM is started with --enable-test-devices.
* src/NetworkManagerDbus.[ch]
- New "getLinkActive" method for devices
- New "setLinkActive" method for devices (only works on test devices)
- New "createTestDevice" method on NetworkManager object to create a test
device of a specified type (ie wired, wireless). UDI is created from
scratch, as is the interface name. Only works when NM is started with
--enable-test-devices switch.
- New "removeTestDevice" method on NetworkManager object which removes a
test device. Only works when NM is started with --enable-test-devices
* src/NetworkManagerDevice.[ch]
- Logic to facilitate test devices. Add variables to NMDevice struct to indicate
whether a device is a test device or not, and what its link status is.
- Deal with test devices in most functions. For those that work directly on hardware
special-case test devices.
- (nm_device_new): don't create a test device if test devices weren't enabled on the
command-line.
- (nm_device_update_link_active): split out logic for wired and wireless device link
checking to separate functions to facilitate test device link checking.
- (nm_device_set_enc_key): Since some drivers for wireless cards are daft and
don't make a distinction between System Authentication and Encryption
(namely Cisco aironet), we use Open System auth when setting a WEP key
on the card. We don't deal with Shared Key auth yet.
- (nm_device_activation_worker): split the activation cancel check logic out into
a separate routine nm_device_activation_cancel_if_needed()
- (nm_device_activation_signal_cancel): rename from nm_device_activation_cancel()
- (nm_device_fake_ap_list): Test wireless devices obviously cannot scan, so create
a list of fake access points that they can "see"
- (nm_device_is_test_device): return whether or not a device is a test device
* src/NetworkManagerPolicy.c
- (nm_policy_get_best_device): attempt to deal with wireless network selection,
previously if you "locked"/forced NM to use a wireless device but then
selected a wireless network for NM to use, it would switch to a wired device.
So, if the active device is wireless and it has a "forced" best AP, use it
if the "forced" best AP is still valid
- (nm_state_modification_monitor): deal with NULL best devices, for example
there were no usable network devices, or the last one was removed
* src/backends/NetworkManager*.c
- Deal with test devices, mostly just return success for operations like getting
a DHCP address
* test/nmtestdevices.c
- Test tool to create/remove/link-switch test devices
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@112 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-31 16:09:15 +00:00
|
|
|
/* Test devices do whatever we tell them to do */
|
|
|
|
|
if (dev->test_device)
|
|
|
|
|
{
|
|
|
|
|
dev->test_device_up = up;
|
|
|
|
|
return;
|
|
|
|
|
}
|
|
|
|
|
|
2004-10-21 19:00:13 +00:00
|
|
|
if (nm_device_get_driver_support_level (dev) == NM_DRIVER_UNSUPPORTED)
|
|
|
|
|
return;
|
|
|
|
|
|
2004-06-24 14:18:37 +00:00
|
|
|
iface_fd = nm_get_network_control_socket ();
|
|
|
|
|
if (iface_fd < 0)
|
|
|
|
|
return;
|
|
|
|
|
|
|
|
|
|
/* Get flags already there */
|
|
|
|
|
strcpy (ifr.ifr_name, nm_device_get_iface (dev));
|
|
|
|
|
err = ioctl (iface_fd, SIOCGIFFLAGS, &ifr);
|
|
|
|
|
if (!err)
|
|
|
|
|
{
|
|
|
|
|
/* If the interface doesn't have those flags already,
|
|
|
|
|
* set them on it.
|
|
|
|
|
*/
|
|
|
|
|
if ((ifr.ifr_flags^flags) & IFF_UP)
|
|
|
|
|
{
|
|
|
|
|
ifr.ifr_flags &= ~IFF_UP;
|
|
|
|
|
ifr.ifr_flags |= IFF_UP & flags;
|
|
|
|
|
err = ioctl (iface_fd, SIOCSIFFLAGS, &ifr);
|
|
|
|
|
if (err)
|
2004-08-23 19:20:49 +00:00
|
|
|
syslog (LOG_ERR, "nm_device_set_up_down() could not bring device %s %s. errno = %d", nm_device_get_iface (dev), (up ? "up" : "down"), errno );
|
2004-06-24 14:18:37 +00:00
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
else
|
2004-08-23 19:20:49 +00:00
|
|
|
syslog (LOG_ERR, "nm_device_set_up_down() could not get flags for device %s. errno = %d", nm_device_get_iface (dev), errno );
|
2004-06-24 14:18:37 +00:00
|
|
|
|
|
|
|
|
close (iface_fd);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* Interface state functions: bring up, down, check
|
|
|
|
|
*
|
|
|
|
|
*/
|
|
|
|
|
void nm_device_bring_up (NMDevice *dev)
|
|
|
|
|
{
|
|
|
|
|
g_return_if_fail (dev != NULL);
|
|
|
|
|
|
|
|
|
|
nm_device_set_up_down (dev, TRUE);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
void nm_device_bring_down (NMDevice *dev)
|
|
|
|
|
{
|
|
|
|
|
g_return_if_fail (dev != NULL);
|
|
|
|
|
|
|
|
|
|
nm_device_set_up_down (dev, FALSE);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
gboolean nm_device_is_up (NMDevice *dev)
|
|
|
|
|
{
|
|
|
|
|
int iface_fd;
|
|
|
|
|
struct ifreq ifr;
|
|
|
|
|
int err;
|
|
|
|
|
|
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
|
|
|
g_return_val_if_fail (dev != NULL, FALSE);
|
2004-06-24 14:18:37 +00:00
|
|
|
|
2004-08-31 Dan Williams <dcbw@redhat.com>
* Remove 'debug' extern global from all files since we now
use syslog()
* src/NetworkManager.[ch]
- Break out routine that get the net.interface property from HAL,
removing that logic from nm_create_device_and_add_to_list()
- (nm_create_device_and_add_to_list): make this a bit more general so
it doesn't do the talking to HAL. Also add arguments to facilitate
the create of test devices.
- (nm_data_mark_state_changed): rename from nm_data_set_state_modified()
- (nm_data_new, main, nm_print_usage): add new argument "--enable-test-devices"
which makes NetworkManager listen for dbus commands to create test
devices, which have no backing hardware. Use when you're on a plane
for example, and/or forgot your wireless card at home. Test devices
_cannot_ be created unless NM is started with --enable-test-devices.
* src/NetworkManagerDbus.[ch]
- New "getLinkActive" method for devices
- New "setLinkActive" method for devices (only works on test devices)
- New "createTestDevice" method on NetworkManager object to create a test
device of a specified type (ie wired, wireless). UDI is created from
scratch, as is the interface name. Only works when NM is started with
--enable-test-devices switch.
- New "removeTestDevice" method on NetworkManager object which removes a
test device. Only works when NM is started with --enable-test-devices
* src/NetworkManagerDevice.[ch]
- Logic to facilitate test devices. Add variables to NMDevice struct to indicate
whether a device is a test device or not, and what its link status is.
- Deal with test devices in most functions. For those that work directly on hardware
special-case test devices.
- (nm_device_new): don't create a test device if test devices weren't enabled on the
command-line.
- (nm_device_update_link_active): split out logic for wired and wireless device link
checking to separate functions to facilitate test device link checking.
- (nm_device_set_enc_key): Since some drivers for wireless cards are daft and
don't make a distinction between System Authentication and Encryption
(namely Cisco aironet), we use Open System auth when setting a WEP key
on the card. We don't deal with Shared Key auth yet.
- (nm_device_activation_worker): split the activation cancel check logic out into
a separate routine nm_device_activation_cancel_if_needed()
- (nm_device_activation_signal_cancel): rename from nm_device_activation_cancel()
- (nm_device_fake_ap_list): Test wireless devices obviously cannot scan, so create
a list of fake access points that they can "see"
- (nm_device_is_test_device): return whether or not a device is a test device
* src/NetworkManagerPolicy.c
- (nm_policy_get_best_device): attempt to deal with wireless network selection,
previously if you "locked"/forced NM to use a wireless device but then
selected a wireless network for NM to use, it would switch to a wired device.
So, if the active device is wireless and it has a "forced" best AP, use it
if the "forced" best AP is still valid
- (nm_state_modification_monitor): deal with NULL best devices, for example
there were no usable network devices, or the last one was removed
* src/backends/NetworkManager*.c
- Deal with test devices, mostly just return success for operations like getting
a DHCP address
* test/nmtestdevices.c
- Test tool to create/remove/link-switch test devices
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@112 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-31 16:09:15 +00:00
|
|
|
if (dev->test_device)
|
|
|
|
|
return (dev->test_device_up);
|
|
|
|
|
|
2004-06-24 14:18:37 +00:00
|
|
|
iface_fd = nm_get_network_control_socket ();
|
|
|
|
|
if (iface_fd < 0)
|
|
|
|
|
return (FALSE);
|
|
|
|
|
|
|
|
|
|
/* Get device's flags */
|
|
|
|
|
strcpy (ifr.ifr_name, nm_device_get_iface (dev));
|
|
|
|
|
err = ioctl (iface_fd, SIOCGIFFLAGS, &ifr);
|
|
|
|
|
close (iface_fd);
|
|
|
|
|
if (!err)
|
|
|
|
|
return (!((ifr.ifr_flags^IFF_UP) & IFF_UP));
|
|
|
|
|
|
2004-08-23 19:20:49 +00:00
|
|
|
syslog (LOG_ERR, "nm_device_is_up() could not get flags for device %s. errno = %d", nm_device_get_iface (dev), errno );
|
2004-06-24 14:18:37 +00:00
|
|
|
return (FALSE);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
2004-10-12 16:44:58 +00:00
|
|
|
/*
|
|
|
|
|
* nm_device_set_mode_managed
|
|
|
|
|
*
|
|
|
|
|
* Set managed/infrastructure mode on a device (currently wireless only)
|
|
|
|
|
*
|
|
|
|
|
*/
|
|
|
|
|
void nm_device_set_mode_managed (NMDevice *dev)
|
|
|
|
|
{
|
|
|
|
|
int sk;
|
|
|
|
|
struct iwreq wreq;
|
|
|
|
|
|
|
|
|
|
g_return_if_fail (dev != NULL);
|
|
|
|
|
g_return_if_fail (nm_device_is_wireless (dev));
|
|
|
|
|
|
|
|
|
|
/* Force the card into Managed/Infrastructure mode */
|
|
|
|
|
sk = iw_sockets_open ();
|
|
|
|
|
if (sk >= 0)
|
|
|
|
|
{
|
|
|
|
|
int err;
|
|
|
|
|
wreq.u.mode = IW_MODE_INFRA;
|
|
|
|
|
err = iw_set_ext (sk, nm_device_get_iface (dev), SIOCSIWMODE, &wreq);
|
|
|
|
|
if (err == -1)
|
|
|
|
|
syslog (LOG_ERR, "nm_device_set_mode_managed (%s): error setting card to Infrastructure mode. errno = %d", nm_device_get_iface (dev), errno);
|
|
|
|
|
close (sk);
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* nm_device_set_mode_adhoc
|
|
|
|
|
*
|
|
|
|
|
* Set Ad Hoc mode on a device (currently wireless only)
|
|
|
|
|
*
|
|
|
|
|
*/
|
|
|
|
|
void nm_device_set_mode_adhoc (NMDevice *dev)
|
|
|
|
|
{
|
|
|
|
|
int sk;
|
|
|
|
|
struct iwreq wreq;
|
|
|
|
|
|
|
|
|
|
g_return_if_fail (dev != NULL);
|
|
|
|
|
g_return_if_fail (nm_device_is_wireless (dev));
|
|
|
|
|
|
2004-10-15 15:59:25 +00:00
|
|
|
/* Force the card into Adhoc mode */
|
2004-10-12 16:44:58 +00:00
|
|
|
sk = iw_sockets_open ();
|
|
|
|
|
if (sk >= 0)
|
|
|
|
|
{
|
|
|
|
|
int err;
|
|
|
|
|
wreq.u.mode = IW_MODE_ADHOC;
|
|
|
|
|
err = iw_set_ext (sk, nm_device_get_iface (dev), SIOCSIWMODE, &wreq);
|
|
|
|
|
if (err == -1)
|
|
|
|
|
syslog (LOG_ERR, "nm_device_set_mode_adhoc (%s): error setting card to Ad Hoc mode. errno = %d", nm_device_get_iface (dev), errno);
|
|
|
|
|
close (sk);
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
2004-07-06 01:34:10 +00:00
|
|
|
/*
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
* nm_device_activation_begin
|
2004-07-06 01:34:10 +00:00
|
|
|
*
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
* Spawn a new thread to handle device activation.
|
|
|
|
|
*
|
|
|
|
|
* Returns: TRUE on success activation beginning
|
|
|
|
|
* FALSE on error beginning activation (bad params, couldn't create thread)
|
2004-07-06 01:34:10 +00:00
|
|
|
*
|
|
|
|
|
*/
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
gboolean nm_device_activation_begin (NMDevice *dev)
|
2004-07-06 01:34:10 +00:00
|
|
|
{
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
GError *error = NULL;
|
2004-08-24 Dan Williams <dcbw@redhat.com>
* src/NetworkManagerAP.[ch]
- Add a "enc_method_good" member and accessors to an Access Point
to signal when we've found the correct encryption method
for an access point
- Add a "timestamp" member and accessors, remove "priority" member
and accessors (use timestamps instead)
- Rename "wep_key"->"enc_key"
- (nm_ap_get_enc_key_hashed): new, return the correct mangled key
for a specified encryption method using the access points
source encryption key/passphrase
* src/NetworkManagerAPList.c
- When updating a network with dbus, grab timestamp now instead of
priority
* src/NetworkManagerDBus.[ch]
- Add signal for "DeviceActivating"
- Switch priority->timestamp
* src/NetworkManagerDevice.c
- Change references of "wep_key" -> "enc_key" or "key"
- Signal DeviceActivating when starting activation
- When activating a wireless device, if the access point we are connecting
to is encrypted, and we have a source key, try to generate a mangled
key and use that (ie, generate real WEP key from a passphrase)
- Rework device activation to fallback to other encryption methods if
a previous one didn't work (ie, try mangling a key as a 104-bit passphrase
first, then if that doesn't work fall back to direct hex key).
- (nm_device_update_best_ap): fix a deadlock, and use timestamps instead of
priority. We now prefer the latest access point used, rather than using
a priority scheme
- (nm_device_do_normal_scan): make the encryption method "unknown" on access
points we've just discovered, and merge in correct info from the global
access point lists
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@68 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-25 22:41:12 +00:00
|
|
|
NMData *data = (NMData *)dev->app_data;
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
|
|
|
|
|
g_return_val_if_fail (dev != NULL, FALSE);
|
2004-10-23 06:43:01 +00:00
|
|
|
g_return_val_if_fail (!dev->activating, TRUE); /* Return if activation has already begun */
|
2004-08-24 Dan Williams <dcbw@redhat.com>
* src/NetworkManagerAP.[ch]
- Add a "enc_method_good" member and accessors to an Access Point
to signal when we've found the correct encryption method
for an access point
- Add a "timestamp" member and accessors, remove "priority" member
and accessors (use timestamps instead)
- Rename "wep_key"->"enc_key"
- (nm_ap_get_enc_key_hashed): new, return the correct mangled key
for a specified encryption method using the access points
source encryption key/passphrase
* src/NetworkManagerAPList.c
- When updating a network with dbus, grab timestamp now instead of
priority
* src/NetworkManagerDBus.[ch]
- Add signal for "DeviceActivating"
- Switch priority->timestamp
* src/NetworkManagerDevice.c
- Change references of "wep_key" -> "enc_key" or "key"
- Signal DeviceActivating when starting activation
- When activating a wireless device, if the access point we are connecting
to is encrypted, and we have a source key, try to generate a mangled
key and use that (ie, generate real WEP key from a passphrase)
- Rework device activation to fallback to other encryption methods if
a previous one didn't work (ie, try mangling a key as a 104-bit passphrase
first, then if that doesn't work fall back to direct hex key).
- (nm_device_update_best_ap): fix a deadlock, and use timestamps instead of
priority. We now prefer the latest access point used, rather than using
a priority scheme
- (nm_device_do_normal_scan): make the encryption method "unknown" on access
points we've just discovered, and merge in correct info from the global
access point lists
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@68 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-25 22:41:12 +00:00
|
|
|
g_return_val_if_fail (data != NULL, FALSE);
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
|
|
|
|
|
/* Ref the device so it doesn't go away while worker function is active */
|
|
|
|
|
nm_device_ref (dev);
|
|
|
|
|
|
2004-09-28 19:50:37 +00:00
|
|
|
/* Don't attempt to actually activate if we are just starting NetworkManager and
|
|
|
|
|
* we are about to activate a wired device that's already configured. Plays nicer
|
|
|
|
|
* with the system when NM is started after a network is already set up.
|
|
|
|
|
*
|
|
|
|
|
* FIXME: IPv6 here too, and this really should not be here, it should be part of
|
|
|
|
|
* the policy, not the device code itself.
|
|
|
|
|
*/
|
|
|
|
|
if (data->starting_up && nm_device_is_wired (data->active_device) && nm_device_get_ip4_address (data->active_device))
|
|
|
|
|
{
|
|
|
|
|
dev->activating = FALSE;
|
|
|
|
|
dev->just_activated = TRUE;
|
|
|
|
|
return (TRUE);
|
|
|
|
|
}
|
|
|
|
|
|
2004-10-21 19:00:13 +00:00
|
|
|
if (nm_device_get_driver_support_level (dev) == NM_DRIVER_UNSUPPORTED)
|
|
|
|
|
{
|
|
|
|
|
dev->activating = FALSE;
|
|
|
|
|
return (FALSE);
|
|
|
|
|
}
|
|
|
|
|
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
/* Reset communication flags between worker and main thread */
|
2004-08-26 21:11:27 +00:00
|
|
|
dev->activating = TRUE;
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
dev->just_activated = FALSE;
|
|
|
|
|
dev->quit_activation = FALSE;
|
2004-10-23 06:43:01 +00:00
|
|
|
dev->activation_failed = FALSE;
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
if (nm_device_is_wireless (dev))
|
2004-10-07 01:50:39 +00:00
|
|
|
{
|
|
|
|
|
dev->options.wireless.now_scanning = FALSE;
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
dev->options.wireless.user_key_received = FALSE;
|
2004-10-07 01:50:39 +00:00
|
|
|
}
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
|
|
|
|
|
if (!g_thread_create (nm_device_activation_worker, dev, FALSE, &error))
|
|
|
|
|
{
|
2004-08-23 19:20:49 +00:00
|
|
|
syslog (LOG_CRIT, "nm_device_activation_begin(): could not create activation worker thread.");
|
2004-08-26 21:11:27 +00:00
|
|
|
dev->activating = FALSE;
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
return (FALSE);
|
|
|
|
|
}
|
|
|
|
|
|
2004-08-26 20:05:24 +00:00
|
|
|
nm_dbus_signal_device_status_change (data->dbus_connection, dev, DEVICE_ACTIVATING);
|
2004-08-24 Dan Williams <dcbw@redhat.com>
* src/NetworkManagerAP.[ch]
- Add a "enc_method_good" member and accessors to an Access Point
to signal when we've found the correct encryption method
for an access point
- Add a "timestamp" member and accessors, remove "priority" member
and accessors (use timestamps instead)
- Rename "wep_key"->"enc_key"
- (nm_ap_get_enc_key_hashed): new, return the correct mangled key
for a specified encryption method using the access points
source encryption key/passphrase
* src/NetworkManagerAPList.c
- When updating a network with dbus, grab timestamp now instead of
priority
* src/NetworkManagerDBus.[ch]
- Add signal for "DeviceActivating"
- Switch priority->timestamp
* src/NetworkManagerDevice.c
- Change references of "wep_key" -> "enc_key" or "key"
- Signal DeviceActivating when starting activation
- When activating a wireless device, if the access point we are connecting
to is encrypted, and we have a source key, try to generate a mangled
key and use that (ie, generate real WEP key from a passphrase)
- Rework device activation to fallback to other encryption methods if
a previous one didn't work (ie, try mangling a key as a 104-bit passphrase
first, then if that doesn't work fall back to direct hex key).
- (nm_device_update_best_ap): fix a deadlock, and use timestamps instead of
priority. We now prefer the latest access point used, rather than using
a priority scheme
- (nm_device_do_normal_scan): make the encryption method "unknown" on access
points we've just discovered, and merge in correct info from the global
access point lists
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@68 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-25 22:41:12 +00:00
|
|
|
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
return (TRUE);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
2004-09-09 20:02:59 +00:00
|
|
|
/*
|
2004-09-12 03:06:44 +00:00
|
|
|
* nm_device_activation_should_cancel
|
2004-09-09 20:02:59 +00:00
|
|
|
*
|
|
|
|
|
* Check whether we should stop activation, and if so clean up flags
|
|
|
|
|
* and other random things.
|
|
|
|
|
*
|
|
|
|
|
*/
|
2004-10-08 05:22:17 +00:00
|
|
|
static gboolean nm_device_activation_should_cancel (NMDevice *dev)
|
2004-09-09 20:02:59 +00:00
|
|
|
{
|
|
|
|
|
g_return_val_if_fail (dev != NULL, TRUE);
|
|
|
|
|
|
|
|
|
|
/* If we were told to quit activation, stop the thread and return */
|
|
|
|
|
if (dev->quit_activation)
|
|
|
|
|
{
|
|
|
|
|
syslog (LOG_DEBUG, "nm_device_activation_worker(%s): activation canceled.", nm_device_get_iface (dev));
|
|
|
|
|
dev->activating = FALSE;
|
|
|
|
|
dev->just_activated = FALSE;
|
2004-10-23 06:43:01 +00:00
|
|
|
dev->activation_failed = TRUE;
|
|
|
|
|
dev->quit_activation = FALSE;
|
2004-09-09 20:02:59 +00:00
|
|
|
return (TRUE);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
return (FALSE);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
/*
|
2004-09-08 18:14:42 +00:00
|
|
|
* nm_device_activate_wireless
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
*
|
|
|
|
|
* Bring up a wireless card with the essid and wep key of its "best" ap
|
|
|
|
|
*
|
|
|
|
|
* Returns: TRUE on successful activation
|
|
|
|
|
* FALSE on unsuccessful activation (ie no best AP)
|
|
|
|
|
*
|
|
|
|
|
*/
|
2004-10-29 16:37:43 +00:00
|
|
|
static gboolean nm_device_activate_wireless (NMDevice *dev, NMAccessPoint *ap, guint *bad_crypt_packets)
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
{
|
2004-07-06 01:34:10 +00:00
|
|
|
gboolean success = FALSE;
|
2004-10-29 16:37:43 +00:00
|
|
|
const char *essid = NULL;
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
|
|
|
|
|
g_return_val_if_fail (dev != NULL, FALSE);
|
|
|
|
|
g_return_val_if_fail (nm_device_is_wireless (dev), FALSE);
|
2004-10-29 16:37:43 +00:00
|
|
|
g_return_val_if_fail (ap != NULL, FALSE);
|
|
|
|
|
g_return_val_if_fail (nm_ap_get_essid (ap) != NULL, FALSE);
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
|
2004-10-08 05:22:17 +00:00
|
|
|
*bad_crypt_packets = 0;
|
2004-08-24 Dan Williams <dcbw@redhat.com>
* src/NetworkManagerAP.[ch]
- Add a "enc_method_good" member and accessors to an Access Point
to signal when we've found the correct encryption method
for an access point
- Add a "timestamp" member and accessors, remove "priority" member
and accessors (use timestamps instead)
- Rename "wep_key"->"enc_key"
- (nm_ap_get_enc_key_hashed): new, return the correct mangled key
for a specified encryption method using the access points
source encryption key/passphrase
* src/NetworkManagerAPList.c
- When updating a network with dbus, grab timestamp now instead of
priority
* src/NetworkManagerDBus.[ch]
- Add signal for "DeviceActivating"
- Switch priority->timestamp
* src/NetworkManagerDevice.c
- Change references of "wep_key" -> "enc_key" or "key"
- Signal DeviceActivating when starting activation
- When activating a wireless device, if the access point we are connecting
to is encrypted, and we have a source key, try to generate a mangled
key and use that (ie, generate real WEP key from a passphrase)
- Rework device activation to fallback to other encryption methods if
a previous one didn't work (ie, try mangling a key as a 104-bit passphrase
first, then if that doesn't work fall back to direct hex key).
- (nm_device_update_best_ap): fix a deadlock, and use timestamps instead of
priority. We now prefer the latest access point used, rather than using
a priority scheme
- (nm_device_do_normal_scan): make the encryption method "unknown" on access
points we've just discovered, and merge in correct info from the global
access point lists
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@68 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-25 22:41:12 +00:00
|
|
|
|
2004-10-29 16:37:43 +00:00
|
|
|
/* Force the card into Managed/Infrastructure mode */
|
|
|
|
|
nm_device_set_mode_managed (dev);
|
2004-10-29 19:52:35 +00:00
|
|
|
nm_device_set_essid (dev, " ");
|
2004-10-12 11:39:08 +00:00
|
|
|
|
2004-10-29 16:37:43 +00:00
|
|
|
/* Disable encryption, then re-enable and set correct key on the card
|
|
|
|
|
* if we are going to encrypt traffic.
|
|
|
|
|
*/
|
|
|
|
|
essid = nm_ap_get_essid (ap);
|
2004-11-03 07:47:32 +00:00
|
|
|
nm_device_set_enc_key (dev, NULL, NM_DEVICE_AUTH_METHOD_NONE);
|
2004-10-29 16:37:43 +00:00
|
|
|
if (nm_ap_get_encrypted (ap) && nm_ap_get_enc_key_source (ap))
|
|
|
|
|
{
|
|
|
|
|
char *hashed_key = nm_ap_get_enc_key_hashed (ap);
|
2004-11-03 07:47:32 +00:00
|
|
|
nm_device_set_enc_key (dev, hashed_key, NM_DEVICE_AUTH_METHOD_SHARED_KEY);
|
2004-10-29 16:37:43 +00:00
|
|
|
g_free (hashed_key);
|
|
|
|
|
}
|
2004-08-24 Dan Williams <dcbw@redhat.com>
* src/NetworkManagerAP.[ch]
- Add a "enc_method_good" member and accessors to an Access Point
to signal when we've found the correct encryption method
for an access point
- Add a "timestamp" member and accessors, remove "priority" member
and accessors (use timestamps instead)
- Rename "wep_key"->"enc_key"
- (nm_ap_get_enc_key_hashed): new, return the correct mangled key
for a specified encryption method using the access points
source encryption key/passphrase
* src/NetworkManagerAPList.c
- When updating a network with dbus, grab timestamp now instead of
priority
* src/NetworkManagerDBus.[ch]
- Add signal for "DeviceActivating"
- Switch priority->timestamp
* src/NetworkManagerDevice.c
- Change references of "wep_key" -> "enc_key" or "key"
- Signal DeviceActivating when starting activation
- When activating a wireless device, if the access point we are connecting
to is encrypted, and we have a source key, try to generate a mangled
key and use that (ie, generate real WEP key from a passphrase)
- Rework device activation to fallback to other encryption methods if
a previous one didn't work (ie, try mangling a key as a 104-bit passphrase
first, then if that doesn't work fall back to direct hex key).
- (nm_device_update_best_ap): fix a deadlock, and use timestamps instead of
priority. We now prefer the latest access point used, rather than using
a priority scheme
- (nm_device_do_normal_scan): make the encryption method "unknown" on access
points we've just discovered, and merge in correct info from the global
access point lists
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@68 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-25 22:41:12 +00:00
|
|
|
|
2004-10-29 16:37:43 +00:00
|
|
|
nm_device_set_essid (dev, essid);
|
|
|
|
|
*bad_crypt_packets = nm_device_get_bad_crypt_packets (dev);
|
2004-08-20 18:18:37 +00:00
|
|
|
|
2004-10-29 16:37:43 +00:00
|
|
|
syslog (LOG_INFO, "nm_device_wireless_activate(%s) using essid '%s'", nm_device_get_iface (dev), essid);
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
|
2004-10-29 16:37:43 +00:00
|
|
|
/* Bring the device up and pause to allow card to associate */
|
|
|
|
|
nm_device_bring_up (dev);
|
|
|
|
|
g_usleep (G_USEC_PER_SEC * 2);
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
|
2004-10-29 16:37:43 +00:00
|
|
|
nm_device_update_link_active (dev, FALSE);
|
|
|
|
|
success = TRUE;
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
|
|
|
|
|
return (success);
|
|
|
|
|
}
|
|
|
|
|
|
2004-08-31 Dan Williams <dcbw@redhat.com>
* Remove 'debug' extern global from all files since we now
use syslog()
* src/NetworkManager.[ch]
- Break out routine that get the net.interface property from HAL,
removing that logic from nm_create_device_and_add_to_list()
- (nm_create_device_and_add_to_list): make this a bit more general so
it doesn't do the talking to HAL. Also add arguments to facilitate
the create of test devices.
- (nm_data_mark_state_changed): rename from nm_data_set_state_modified()
- (nm_data_new, main, nm_print_usage): add new argument "--enable-test-devices"
which makes NetworkManager listen for dbus commands to create test
devices, which have no backing hardware. Use when you're on a plane
for example, and/or forgot your wireless card at home. Test devices
_cannot_ be created unless NM is started with --enable-test-devices.
* src/NetworkManagerDbus.[ch]
- New "getLinkActive" method for devices
- New "setLinkActive" method for devices (only works on test devices)
- New "createTestDevice" method on NetworkManager object to create a test
device of a specified type (ie wired, wireless). UDI is created from
scratch, as is the interface name. Only works when NM is started with
--enable-test-devices switch.
- New "removeTestDevice" method on NetworkManager object which removes a
test device. Only works when NM is started with --enable-test-devices
* src/NetworkManagerDevice.[ch]
- Logic to facilitate test devices. Add variables to NMDevice struct to indicate
whether a device is a test device or not, and what its link status is.
- Deal with test devices in most functions. For those that work directly on hardware
special-case test devices.
- (nm_device_new): don't create a test device if test devices weren't enabled on the
command-line.
- (nm_device_update_link_active): split out logic for wired and wireless device link
checking to separate functions to facilitate test device link checking.
- (nm_device_set_enc_key): Since some drivers for wireless cards are daft and
don't make a distinction between System Authentication and Encryption
(namely Cisco aironet), we use Open System auth when setting a WEP key
on the card. We don't deal with Shared Key auth yet.
- (nm_device_activation_worker): split the activation cancel check logic out into
a separate routine nm_device_activation_cancel_if_needed()
- (nm_device_activation_signal_cancel): rename from nm_device_activation_cancel()
- (nm_device_fake_ap_list): Test wireless devices obviously cannot scan, so create
a list of fake access points that they can "see"
- (nm_device_is_test_device): return whether or not a device is a test device
* src/NetworkManagerPolicy.c
- (nm_policy_get_best_device): attempt to deal with wireless network selection,
previously if you "locked"/forced NM to use a wireless device but then
selected a wireless network for NM to use, it would switch to a wired device.
So, if the active device is wireless and it has a "forced" best AP, use it
if the "forced" best AP is still valid
- (nm_state_modification_monitor): deal with NULL best devices, for example
there were no usable network devices, or the last one was removed
* src/backends/NetworkManager*.c
- Deal with test devices, mostly just return success for operations like getting
a DHCP address
* test/nmtestdevices.c
- Test tool to create/remove/link-switch test devices
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@112 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-31 16:09:15 +00:00
|
|
|
|
2004-10-29 16:37:43 +00:00
|
|
|
gboolean AP_NEED_KEY (NMAccessPoint *ap)
|
|
|
|
|
{
|
|
|
|
|
char *s;
|
|
|
|
|
int len = 0;
|
|
|
|
|
g_return_val_if_fail (ap != NULL, FALSE);
|
|
|
|
|
|
|
|
|
|
if ((s = nm_ap_get_enc_key_source (ap)))
|
|
|
|
|
len = strlen (s);
|
|
|
|
|
|
|
|
|
|
syslog (LOG_NOTICE, "AP_NEED_KEY: is_enc=%d && (!(key_source=%d) || !(strlen=%d))\n", nm_ap_get_encrypted (ap),
|
|
|
|
|
!!nm_ap_get_enc_key_source (ap), len);
|
|
|
|
|
if (nm_ap_get_encrypted (ap) && (!nm_ap_get_enc_key_source (ap) || !len))
|
|
|
|
|
return (TRUE);
|
|
|
|
|
|
|
|
|
|
return (FALSE);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
gboolean HAVE_LINK (NMDevice *dev, guint32 bad_crypt_packets)
|
2004-10-04 15:44:00 +00:00
|
|
|
{
|
|
|
|
|
g_return_val_if_fail (dev != NULL, FALSE);
|
|
|
|
|
g_return_val_if_fail (nm_device_is_wireless (dev), FALSE);
|
|
|
|
|
|
2004-10-14 20:31:35 +00:00
|
|
|
syslog (LOG_NOTICE, "HAVELINK: act=%d && (dev_crypt=%d <= prev_crypt=%d)\n", nm_device_get_link_active (dev), nm_device_get_bad_crypt_packets (dev), bad_crypt_packets);
|
2004-10-04 15:44:00 +00:00
|
|
|
return (nm_device_get_link_active (dev) && (nm_device_get_bad_crypt_packets (dev) <= bad_crypt_packets));
|
|
|
|
|
}
|
|
|
|
|
|
2004-08-31 Dan Williams <dcbw@redhat.com>
* Remove 'debug' extern global from all files since we now
use syslog()
* src/NetworkManager.[ch]
- Break out routine that get the net.interface property from HAL,
removing that logic from nm_create_device_and_add_to_list()
- (nm_create_device_and_add_to_list): make this a bit more general so
it doesn't do the talking to HAL. Also add arguments to facilitate
the create of test devices.
- (nm_data_mark_state_changed): rename from nm_data_set_state_modified()
- (nm_data_new, main, nm_print_usage): add new argument "--enable-test-devices"
which makes NetworkManager listen for dbus commands to create test
devices, which have no backing hardware. Use when you're on a plane
for example, and/or forgot your wireless card at home. Test devices
_cannot_ be created unless NM is started with --enable-test-devices.
* src/NetworkManagerDbus.[ch]
- New "getLinkActive" method for devices
- New "setLinkActive" method for devices (only works on test devices)
- New "createTestDevice" method on NetworkManager object to create a test
device of a specified type (ie wired, wireless). UDI is created from
scratch, as is the interface name. Only works when NM is started with
--enable-test-devices switch.
- New "removeTestDevice" method on NetworkManager object which removes a
test device. Only works when NM is started with --enable-test-devices
* src/NetworkManagerDevice.[ch]
- Logic to facilitate test devices. Add variables to NMDevice struct to indicate
whether a device is a test device or not, and what its link status is.
- Deal with test devices in most functions. For those that work directly on hardware
special-case test devices.
- (nm_device_new): don't create a test device if test devices weren't enabled on the
command-line.
- (nm_device_update_link_active): split out logic for wired and wireless device link
checking to separate functions to facilitate test device link checking.
- (nm_device_set_enc_key): Since some drivers for wireless cards are daft and
don't make a distinction between System Authentication and Encryption
(namely Cisco aironet), we use Open System auth when setting a WEP key
on the card. We don't deal with Shared Key auth yet.
- (nm_device_activation_worker): split the activation cancel check logic out into
a separate routine nm_device_activation_cancel_if_needed()
- (nm_device_activation_signal_cancel): rename from nm_device_activation_cancel()
- (nm_device_fake_ap_list): Test wireless devices obviously cannot scan, so create
a list of fake access points that they can "see"
- (nm_device_is_test_device): return whether or not a device is a test device
* src/NetworkManagerPolicy.c
- (nm_policy_get_best_device): attempt to deal with wireless network selection,
previously if you "locked"/forced NM to use a wireless device but then
selected a wireless network for NM to use, it would switch to a wired device.
So, if the active device is wireless and it has a "forced" best AP, use it
if the "forced" best AP is still valid
- (nm_state_modification_monitor): deal with NULL best devices, for example
there were no usable network devices, or the last one was removed
* src/backends/NetworkManager*.c
- Deal with test devices, mostly just return success for operations like getting
a DHCP address
* test/nmtestdevices.c
- Test tool to create/remove/link-switch test devices
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@112 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-31 16:09:15 +00:00
|
|
|
/*
|
2004-09-09 20:02:59 +00:00
|
|
|
* nm_device_activate_wireless_wait_for_link
|
2004-08-31 Dan Williams <dcbw@redhat.com>
* Remove 'debug' extern global from all files since we now
use syslog()
* src/NetworkManager.[ch]
- Break out routine that get the net.interface property from HAL,
removing that logic from nm_create_device_and_add_to_list()
- (nm_create_device_and_add_to_list): make this a bit more general so
it doesn't do the talking to HAL. Also add arguments to facilitate
the create of test devices.
- (nm_data_mark_state_changed): rename from nm_data_set_state_modified()
- (nm_data_new, main, nm_print_usage): add new argument "--enable-test-devices"
which makes NetworkManager listen for dbus commands to create test
devices, which have no backing hardware. Use when you're on a plane
for example, and/or forgot your wireless card at home. Test devices
_cannot_ be created unless NM is started with --enable-test-devices.
* src/NetworkManagerDbus.[ch]
- New "getLinkActive" method for devices
- New "setLinkActive" method for devices (only works on test devices)
- New "createTestDevice" method on NetworkManager object to create a test
device of a specified type (ie wired, wireless). UDI is created from
scratch, as is the interface name. Only works when NM is started with
--enable-test-devices switch.
- New "removeTestDevice" method on NetworkManager object which removes a
test device. Only works when NM is started with --enable-test-devices
* src/NetworkManagerDevice.[ch]
- Logic to facilitate test devices. Add variables to NMDevice struct to indicate
whether a device is a test device or not, and what its link status is.
- Deal with test devices in most functions. For those that work directly on hardware
special-case test devices.
- (nm_device_new): don't create a test device if test devices weren't enabled on the
command-line.
- (nm_device_update_link_active): split out logic for wired and wireless device link
checking to separate functions to facilitate test device link checking.
- (nm_device_set_enc_key): Since some drivers for wireless cards are daft and
don't make a distinction between System Authentication and Encryption
(namely Cisco aironet), we use Open System auth when setting a WEP key
on the card. We don't deal with Shared Key auth yet.
- (nm_device_activation_worker): split the activation cancel check logic out into
a separate routine nm_device_activation_cancel_if_needed()
- (nm_device_activation_signal_cancel): rename from nm_device_activation_cancel()
- (nm_device_fake_ap_list): Test wireless devices obviously cannot scan, so create
a list of fake access points that they can "see"
- (nm_device_is_test_device): return whether or not a device is a test device
* src/NetworkManagerPolicy.c
- (nm_policy_get_best_device): attempt to deal with wireless network selection,
previously if you "locked"/forced NM to use a wireless device but then
selected a wireless network for NM to use, it would switch to a wired device.
So, if the active device is wireless and it has a "forced" best AP, use it
if the "forced" best AP is still valid
- (nm_state_modification_monitor): deal with NULL best devices, for example
there were no usable network devices, or the last one was removed
* src/backends/NetworkManager*.c
- Deal with test devices, mostly just return success for operations like getting
a DHCP address
* test/nmtestdevices.c
- Test tool to create/remove/link-switch test devices
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@112 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-31 16:09:15 +00:00
|
|
|
*
|
2004-09-09 20:02:59 +00:00
|
|
|
* Spin until we have a wireless link, which may mean
|
|
|
|
|
* requesting a key from the user and trying various hashed
|
|
|
|
|
* iterations of that key.
|
2004-08-31 Dan Williams <dcbw@redhat.com>
* Remove 'debug' extern global from all files since we now
use syslog()
* src/NetworkManager.[ch]
- Break out routine that get the net.interface property from HAL,
removing that logic from nm_create_device_and_add_to_list()
- (nm_create_device_and_add_to_list): make this a bit more general so
it doesn't do the talking to HAL. Also add arguments to facilitate
the create of test devices.
- (nm_data_mark_state_changed): rename from nm_data_set_state_modified()
- (nm_data_new, main, nm_print_usage): add new argument "--enable-test-devices"
which makes NetworkManager listen for dbus commands to create test
devices, which have no backing hardware. Use when you're on a plane
for example, and/or forgot your wireless card at home. Test devices
_cannot_ be created unless NM is started with --enable-test-devices.
* src/NetworkManagerDbus.[ch]
- New "getLinkActive" method for devices
- New "setLinkActive" method for devices (only works on test devices)
- New "createTestDevice" method on NetworkManager object to create a test
device of a specified type (ie wired, wireless). UDI is created from
scratch, as is the interface name. Only works when NM is started with
--enable-test-devices switch.
- New "removeTestDevice" method on NetworkManager object which removes a
test device. Only works when NM is started with --enable-test-devices
* src/NetworkManagerDevice.[ch]
- Logic to facilitate test devices. Add variables to NMDevice struct to indicate
whether a device is a test device or not, and what its link status is.
- Deal with test devices in most functions. For those that work directly on hardware
special-case test devices.
- (nm_device_new): don't create a test device if test devices weren't enabled on the
command-line.
- (nm_device_update_link_active): split out logic for wired and wireless device link
checking to separate functions to facilitate test device link checking.
- (nm_device_set_enc_key): Since some drivers for wireless cards are daft and
don't make a distinction between System Authentication and Encryption
(namely Cisco aironet), we use Open System auth when setting a WEP key
on the card. We don't deal with Shared Key auth yet.
- (nm_device_activation_worker): split the activation cancel check logic out into
a separate routine nm_device_activation_cancel_if_needed()
- (nm_device_activation_signal_cancel): rename from nm_device_activation_cancel()
- (nm_device_fake_ap_list): Test wireless devices obviously cannot scan, so create
a list of fake access points that they can "see"
- (nm_device_is_test_device): return whether or not a device is a test device
* src/NetworkManagerPolicy.c
- (nm_policy_get_best_device): attempt to deal with wireless network selection,
previously if you "locked"/forced NM to use a wireless device but then
selected a wireless network for NM to use, it would switch to a wired device.
So, if the active device is wireless and it has a "forced" best AP, use it
if the "forced" best AP is still valid
- (nm_state_modification_monitor): deal with NULL best devices, for example
there were no usable network devices, or the last one was removed
* src/backends/NetworkManager*.c
- Deal with test devices, mostly just return success for operations like getting
a DHCP address
* test/nmtestdevices.c
- Test tool to create/remove/link-switch test devices
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@112 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-31 16:09:15 +00:00
|
|
|
*
|
|
|
|
|
*/
|
2004-09-09 20:02:59 +00:00
|
|
|
void nm_device_activate_wireless_wait_for_link (NMDevice *dev)
|
2004-08-31 Dan Williams <dcbw@redhat.com>
* Remove 'debug' extern global from all files since we now
use syslog()
* src/NetworkManager.[ch]
- Break out routine that get the net.interface property from HAL,
removing that logic from nm_create_device_and_add_to_list()
- (nm_create_device_and_add_to_list): make this a bit more general so
it doesn't do the talking to HAL. Also add arguments to facilitate
the create of test devices.
- (nm_data_mark_state_changed): rename from nm_data_set_state_modified()
- (nm_data_new, main, nm_print_usage): add new argument "--enable-test-devices"
which makes NetworkManager listen for dbus commands to create test
devices, which have no backing hardware. Use when you're on a plane
for example, and/or forgot your wireless card at home. Test devices
_cannot_ be created unless NM is started with --enable-test-devices.
* src/NetworkManagerDbus.[ch]
- New "getLinkActive" method for devices
- New "setLinkActive" method for devices (only works on test devices)
- New "createTestDevice" method on NetworkManager object to create a test
device of a specified type (ie wired, wireless). UDI is created from
scratch, as is the interface name. Only works when NM is started with
--enable-test-devices switch.
- New "removeTestDevice" method on NetworkManager object which removes a
test device. Only works when NM is started with --enable-test-devices
* src/NetworkManagerDevice.[ch]
- Logic to facilitate test devices. Add variables to NMDevice struct to indicate
whether a device is a test device or not, and what its link status is.
- Deal with test devices in most functions. For those that work directly on hardware
special-case test devices.
- (nm_device_new): don't create a test device if test devices weren't enabled on the
command-line.
- (nm_device_update_link_active): split out logic for wired and wireless device link
checking to separate functions to facilitate test device link checking.
- (nm_device_set_enc_key): Since some drivers for wireless cards are daft and
don't make a distinction between System Authentication and Encryption
(namely Cisco aironet), we use Open System auth when setting a WEP key
on the card. We don't deal with Shared Key auth yet.
- (nm_device_activation_worker): split the activation cancel check logic out into
a separate routine nm_device_activation_cancel_if_needed()
- (nm_device_activation_signal_cancel): rename from nm_device_activation_cancel()
- (nm_device_fake_ap_list): Test wireless devices obviously cannot scan, so create
a list of fake access points that they can "see"
- (nm_device_is_test_device): return whether or not a device is a test device
* src/NetworkManagerPolicy.c
- (nm_policy_get_best_device): attempt to deal with wireless network selection,
previously if you "locked"/forced NM to use a wireless device but then
selected a wireless network for NM to use, it would switch to a wired device.
So, if the active device is wireless and it has a "forced" best AP, use it
if the "forced" best AP is still valid
- (nm_state_modification_monitor): deal with NULL best devices, for example
there were no usable network devices, or the last one was removed
* src/backends/NetworkManager*.c
- Deal with test devices, mostly just return success for operations like getting
a DHCP address
* test/nmtestdevices.c
- Test tool to create/remove/link-switch test devices
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@112 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-31 16:09:15 +00:00
|
|
|
{
|
2004-09-09 20:02:59 +00:00
|
|
|
NMAccessPoint *best_ap;
|
2004-10-04 15:44:00 +00:00
|
|
|
guint32 bad_crypt_packets = 0;
|
2004-08-31 Dan Williams <dcbw@redhat.com>
* Remove 'debug' extern global from all files since we now
use syslog()
* src/NetworkManager.[ch]
- Break out routine that get the net.interface property from HAL,
removing that logic from nm_create_device_and_add_to_list()
- (nm_create_device_and_add_to_list): make this a bit more general so
it doesn't do the talking to HAL. Also add arguments to facilitate
the create of test devices.
- (nm_data_mark_state_changed): rename from nm_data_set_state_modified()
- (nm_data_new, main, nm_print_usage): add new argument "--enable-test-devices"
which makes NetworkManager listen for dbus commands to create test
devices, which have no backing hardware. Use when you're on a plane
for example, and/or forgot your wireless card at home. Test devices
_cannot_ be created unless NM is started with --enable-test-devices.
* src/NetworkManagerDbus.[ch]
- New "getLinkActive" method for devices
- New "setLinkActive" method for devices (only works on test devices)
- New "createTestDevice" method on NetworkManager object to create a test
device of a specified type (ie wired, wireless). UDI is created from
scratch, as is the interface name. Only works when NM is started with
--enable-test-devices switch.
- New "removeTestDevice" method on NetworkManager object which removes a
test device. Only works when NM is started with --enable-test-devices
* src/NetworkManagerDevice.[ch]
- Logic to facilitate test devices. Add variables to NMDevice struct to indicate
whether a device is a test device or not, and what its link status is.
- Deal with test devices in most functions. For those that work directly on hardware
special-case test devices.
- (nm_device_new): don't create a test device if test devices weren't enabled on the
command-line.
- (nm_device_update_link_active): split out logic for wired and wireless device link
checking to separate functions to facilitate test device link checking.
- (nm_device_set_enc_key): Since some drivers for wireless cards are daft and
don't make a distinction between System Authentication and Encryption
(namely Cisco aironet), we use Open System auth when setting a WEP key
on the card. We don't deal with Shared Key auth yet.
- (nm_device_activation_worker): split the activation cancel check logic out into
a separate routine nm_device_activation_cancel_if_needed()
- (nm_device_activation_signal_cancel): rename from nm_device_activation_cancel()
- (nm_device_fake_ap_list): Test wireless devices obviously cannot scan, so create
a list of fake access points that they can "see"
- (nm_device_is_test_device): return whether or not a device is a test device
* src/NetworkManagerPolicy.c
- (nm_policy_get_best_device): attempt to deal with wireless network selection,
previously if you "locked"/forced NM to use a wireless device but then
selected a wireless network for NM to use, it would switch to a wired device.
So, if the active device is wireless and it has a "forced" best AP, use it
if the "forced" best AP is still valid
- (nm_state_modification_monitor): deal with NULL best devices, for example
there were no usable network devices, or the last one was removed
* src/backends/NetworkManager*.c
- Deal with test devices, mostly just return success for operations like getting
a DHCP address
* test/nmtestdevices.c
- Test tool to create/remove/link-switch test devices
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@112 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-31 16:09:15 +00:00
|
|
|
|
2004-09-09 20:02:59 +00:00
|
|
|
g_return_if_fail (dev != NULL);
|
|
|
|
|
|
|
|
|
|
/* If the card is just inserted, we may not have had a chance to scan yet */
|
2004-10-29 16:37:43 +00:00
|
|
|
best_ap = nm_device_get_best_ap (dev);
|
|
|
|
|
if (!best_ap)
|
2004-08-31 Dan Williams <dcbw@redhat.com>
* Remove 'debug' extern global from all files since we now
use syslog()
* src/NetworkManager.[ch]
- Break out routine that get the net.interface property from HAL,
removing that logic from nm_create_device_and_add_to_list()
- (nm_create_device_and_add_to_list): make this a bit more general so
it doesn't do the talking to HAL. Also add arguments to facilitate
the create of test devices.
- (nm_data_mark_state_changed): rename from nm_data_set_state_modified()
- (nm_data_new, main, nm_print_usage): add new argument "--enable-test-devices"
which makes NetworkManager listen for dbus commands to create test
devices, which have no backing hardware. Use when you're on a plane
for example, and/or forgot your wireless card at home. Test devices
_cannot_ be created unless NM is started with --enable-test-devices.
* src/NetworkManagerDbus.[ch]
- New "getLinkActive" method for devices
- New "setLinkActive" method for devices (only works on test devices)
- New "createTestDevice" method on NetworkManager object to create a test
device of a specified type (ie wired, wireless). UDI is created from
scratch, as is the interface name. Only works when NM is started with
--enable-test-devices switch.
- New "removeTestDevice" method on NetworkManager object which removes a
test device. Only works when NM is started with --enable-test-devices
* src/NetworkManagerDevice.[ch]
- Logic to facilitate test devices. Add variables to NMDevice struct to indicate
whether a device is a test device or not, and what its link status is.
- Deal with test devices in most functions. For those that work directly on hardware
special-case test devices.
- (nm_device_new): don't create a test device if test devices weren't enabled on the
command-line.
- (nm_device_update_link_active): split out logic for wired and wireless device link
checking to separate functions to facilitate test device link checking.
- (nm_device_set_enc_key): Since some drivers for wireless cards are daft and
don't make a distinction between System Authentication and Encryption
(namely Cisco aironet), we use Open System auth when setting a WEP key
on the card. We don't deal with Shared Key auth yet.
- (nm_device_activation_worker): split the activation cancel check logic out into
a separate routine nm_device_activation_cancel_if_needed()
- (nm_device_activation_signal_cancel): rename from nm_device_activation_cancel()
- (nm_device_fake_ap_list): Test wireless devices obviously cannot scan, so create
a list of fake access points that they can "see"
- (nm_device_is_test_device): return whether or not a device is a test device
* src/NetworkManagerPolicy.c
- (nm_policy_get_best_device): attempt to deal with wireless network selection,
previously if you "locked"/forced NM to use a wireless device but then
selected a wireless network for NM to use, it would switch to a wired device.
So, if the active device is wireless and it has a "forced" best AP, use it
if the "forced" best AP is still valid
- (nm_state_modification_monitor): deal with NULL best devices, for example
there were no usable network devices, or the last one was removed
* src/backends/NetworkManager*.c
- Deal with test devices, mostly just return success for operations like getting
a DHCP address
* test/nmtestdevices.c
- Test tool to create/remove/link-switch test devices
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@112 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-31 16:09:15 +00:00
|
|
|
{
|
2004-09-09 20:02:59 +00:00
|
|
|
nm_device_do_wireless_scan (dev);
|
2004-10-08 05:22:17 +00:00
|
|
|
nm_device_update_best_ap (dev);
|
2004-09-09 20:02:59 +00:00
|
|
|
best_ap = nm_device_get_best_ap (dev);
|
2004-08-31 Dan Williams <dcbw@redhat.com>
* Remove 'debug' extern global from all files since we now
use syslog()
* src/NetworkManager.[ch]
- Break out routine that get the net.interface property from HAL,
removing that logic from nm_create_device_and_add_to_list()
- (nm_create_device_and_add_to_list): make this a bit more general so
it doesn't do the talking to HAL. Also add arguments to facilitate
the create of test devices.
- (nm_data_mark_state_changed): rename from nm_data_set_state_modified()
- (nm_data_new, main, nm_print_usage): add new argument "--enable-test-devices"
which makes NetworkManager listen for dbus commands to create test
devices, which have no backing hardware. Use when you're on a plane
for example, and/or forgot your wireless card at home. Test devices
_cannot_ be created unless NM is started with --enable-test-devices.
* src/NetworkManagerDbus.[ch]
- New "getLinkActive" method for devices
- New "setLinkActive" method for devices (only works on test devices)
- New "createTestDevice" method on NetworkManager object to create a test
device of a specified type (ie wired, wireless). UDI is created from
scratch, as is the interface name. Only works when NM is started with
--enable-test-devices switch.
- New "removeTestDevice" method on NetworkManager object which removes a
test device. Only works when NM is started with --enable-test-devices
* src/NetworkManagerDevice.[ch]
- Logic to facilitate test devices. Add variables to NMDevice struct to indicate
whether a device is a test device or not, and what its link status is.
- Deal with test devices in most functions. For those that work directly on hardware
special-case test devices.
- (nm_device_new): don't create a test device if test devices weren't enabled on the
command-line.
- (nm_device_update_link_active): split out logic for wired and wireless device link
checking to separate functions to facilitate test device link checking.
- (nm_device_set_enc_key): Since some drivers for wireless cards are daft and
don't make a distinction between System Authentication and Encryption
(namely Cisco aironet), we use Open System auth when setting a WEP key
on the card. We don't deal with Shared Key auth yet.
- (nm_device_activation_worker): split the activation cancel check logic out into
a separate routine nm_device_activation_cancel_if_needed()
- (nm_device_activation_signal_cancel): rename from nm_device_activation_cancel()
- (nm_device_fake_ap_list): Test wireless devices obviously cannot scan, so create
a list of fake access points that they can "see"
- (nm_device_is_test_device): return whether or not a device is a test device
* src/NetworkManagerPolicy.c
- (nm_policy_get_best_device): attempt to deal with wireless network selection,
previously if you "locked"/forced NM to use a wireless device but then
selected a wireless network for NM to use, it would switch to a wired device.
So, if the active device is wireless and it has a "forced" best AP, use it
if the "forced" best AP is still valid
- (nm_state_modification_monitor): deal with NULL best devices, for example
there were no usable network devices, or the last one was removed
* src/backends/NetworkManager*.c
- Deal with test devices, mostly just return success for operations like getting
a DHCP address
* test/nmtestdevices.c
- Test tool to create/remove/link-switch test devices
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@112 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-31 16:09:15 +00:00
|
|
|
}
|
|
|
|
|
|
2004-09-09 20:02:59 +00:00
|
|
|
/* Wait until we have a link. Some things that might block us from
|
|
|
|
|
* getting one:
|
|
|
|
|
* 1) Access point we want to associate with has encryption enabled and
|
|
|
|
|
* we don't have the right encryption key. If we have a key of some
|
|
|
|
|
* sort, try various passhprase->key hashes of it. If we don't have
|
|
|
|
|
* a key, ask the user for one and wait until we are canceled (wireless
|
|
|
|
|
* card was ejected or the user plugged the computer into a wired network)
|
|
|
|
|
* or until we get a key back.
|
|
|
|
|
* 2) We don't have any access points we wish to associate with yet. In that case
|
|
|
|
|
* wait for the wireless scan to complete in the other thread and to pick
|
|
|
|
|
* a "best" access point for us.
|
|
|
|
|
*
|
|
|
|
|
*/
|
2004-10-04 15:44:00 +00:00
|
|
|
|
|
|
|
|
/* There are two ways to check for a good link. If we are using WEP and Open System
|
|
|
|
|
* authentication, then we can associate with the base station regardless of whether the
|
|
|
|
|
* WEP key is right or not. Therefore, we have to monitor the # of packets the card discards
|
|
|
|
|
* when its unable to decrypt them, since that gives us some indicator of whether the WEP
|
|
|
|
|
* key is wrong. It seems that right after association, at least one packet is dropped by
|
|
|
|
|
* most cards if the WEP key is wrong.
|
|
|
|
|
*
|
|
|
|
|
* The second and better way (if all cards actually supported it) is to check the MAC address
|
|
|
|
|
* the card is associated with. However, this doesn't tell us if the WEP key is wrong when we
|
|
|
|
|
* are using Open System authentication. Also, not all drivers return an invalid MAC address
|
|
|
|
|
* when the card cannot communicate with the access point.
|
|
|
|
|
*/
|
2004-10-08 16:07:27 +00:00
|
|
|
|
2004-10-29 16:37:43 +00:00
|
|
|
/* Try activating the device with the key and access point we have already */
|
|
|
|
|
if (best_ap)
|
|
|
|
|
nm_device_activate_wireless (dev, best_ap, &bad_crypt_packets);
|
|
|
|
|
|
2004-10-08 16:07:27 +00:00
|
|
|
/* For the link check, ensure that:
|
|
|
|
|
* 1) a classic link check is good, ie does the card report a valid associated AP MAC address and is it
|
|
|
|
|
* receiving WEP-enabled packets OK if WEP is on
|
|
|
|
|
* 2) we have a best access point, and if that AP is encrypted, that we have a valid encryption key for it
|
|
|
|
|
*
|
|
|
|
|
* If either of these things fail, we try other access points or we try to retrieve another encrpytion key
|
|
|
|
|
* from the user.
|
|
|
|
|
*
|
|
|
|
|
*/
|
2004-10-29 16:37:43 +00:00
|
|
|
while (!HAVE_LINK (dev, bad_crypt_packets) || (best_ap && AP_NEED_KEY (best_ap)))
|
2004-09-09 20:02:59 +00:00
|
|
|
{
|
2004-10-29 16:37:43 +00:00
|
|
|
/* Refresh what we think is the best AP to associate with */
|
2004-10-21 19:33:41 +00:00
|
|
|
if (best_ap)
|
2004-10-29 16:37:43 +00:00
|
|
|
nm_ap_unref (best_ap);
|
2004-09-09 20:02:59 +00:00
|
|
|
if ((best_ap = nm_device_get_best_ap (dev)))
|
|
|
|
|
{
|
2004-10-07 01:50:39 +00:00
|
|
|
dev->options.wireless.now_scanning = FALSE;
|
|
|
|
|
|
2004-10-08 05:22:17 +00:00
|
|
|
/* If we don't have a link yet, the encryption key is bad. Ask the user for a
|
|
|
|
|
* new one.
|
2004-09-09 20:02:59 +00:00
|
|
|
*/
|
2004-10-08 05:22:17 +00:00
|
|
|
if (nm_ap_get_encrypted (best_ap))
|
2004-09-09 20:02:59 +00:00
|
|
|
{
|
2004-10-08 05:22:17 +00:00
|
|
|
dev->options.wireless.user_key_received = FALSE;
|
|
|
|
|
nm_dbus_get_user_key_for_network (dev->app_data->dbus_connection, dev, best_ap);
|
2004-09-09 20:02:59 +00:00
|
|
|
|
2004-10-08 05:22:17 +00:00
|
|
|
/* Wait for the key to come back */
|
|
|
|
|
syslog (LOG_DEBUG, "nm_device_activation_worker(%s): asking for user key.", nm_device_get_iface (dev));
|
|
|
|
|
while (!dev->options.wireless.user_key_received && !dev->quit_activation)
|
|
|
|
|
g_usleep (G_USEC_PER_SEC / 2);
|
2004-11-03 07:47:32 +00:00
|
|
|
|
2004-10-08 05:22:17 +00:00
|
|
|
syslog (LOG_DEBUG, "nm_device_activation_worker(%s): user key received.", nm_device_get_iface (dev));
|
2004-09-09 20:02:59 +00:00
|
|
|
|
|
|
|
|
/* If we were told to quit activation, stop the thread and return */
|
2004-09-12 03:06:44 +00:00
|
|
|
if (nm_device_activation_should_cancel (dev))
|
2004-10-29 16:37:43 +00:00
|
|
|
goto out;
|
2004-09-09 20:02:59 +00:00
|
|
|
}
|
|
|
|
|
|
2004-11-03 07:47:32 +00:00
|
|
|
/* Try activating again with up-to-date access point and keys */
|
2004-10-29 16:37:43 +00:00
|
|
|
nm_device_activate_wireless (dev, best_ap, &bad_crypt_packets);
|
2004-09-09 20:02:59 +00:00
|
|
|
}
|
|
|
|
|
else
|
|
|
|
|
{
|
2004-10-07 01:50:39 +00:00
|
|
|
dev->options.wireless.now_scanning = TRUE;
|
2004-09-09 20:02:59 +00:00
|
|
|
syslog (LOG_DEBUG, "nm_device_activation_worker(%s): waiting for an access point.", nm_device_get_iface (dev));
|
|
|
|
|
g_usleep (G_USEC_PER_SEC * 2);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* If we were told to quit activation, stop the thread and return */
|
2004-09-12 03:06:44 +00:00
|
|
|
if (nm_device_activation_should_cancel (dev))
|
2004-10-29 16:37:43 +00:00
|
|
|
goto out;
|
2004-09-09 20:02:59 +00:00
|
|
|
}
|
2004-10-07 01:50:39 +00:00
|
|
|
|
2004-10-29 16:37:43 +00:00
|
|
|
out:
|
|
|
|
|
if (best_ap)
|
|
|
|
|
nm_ap_unref (best_ap);
|
2004-10-07 01:50:39 +00:00
|
|
|
dev->options.wireless.now_scanning = FALSE;
|
2004-08-31 Dan Williams <dcbw@redhat.com>
* Remove 'debug' extern global from all files since we now
use syslog()
* src/NetworkManager.[ch]
- Break out routine that get the net.interface property from HAL,
removing that logic from nm_create_device_and_add_to_list()
- (nm_create_device_and_add_to_list): make this a bit more general so
it doesn't do the talking to HAL. Also add arguments to facilitate
the create of test devices.
- (nm_data_mark_state_changed): rename from nm_data_set_state_modified()
- (nm_data_new, main, nm_print_usage): add new argument "--enable-test-devices"
which makes NetworkManager listen for dbus commands to create test
devices, which have no backing hardware. Use when you're on a plane
for example, and/or forgot your wireless card at home. Test devices
_cannot_ be created unless NM is started with --enable-test-devices.
* src/NetworkManagerDbus.[ch]
- New "getLinkActive" method for devices
- New "setLinkActive" method for devices (only works on test devices)
- New "createTestDevice" method on NetworkManager object to create a test
device of a specified type (ie wired, wireless). UDI is created from
scratch, as is the interface name. Only works when NM is started with
--enable-test-devices switch.
- New "removeTestDevice" method on NetworkManager object which removes a
test device. Only works when NM is started with --enable-test-devices
* src/NetworkManagerDevice.[ch]
- Logic to facilitate test devices. Add variables to NMDevice struct to indicate
whether a device is a test device or not, and what its link status is.
- Deal with test devices in most functions. For those that work directly on hardware
special-case test devices.
- (nm_device_new): don't create a test device if test devices weren't enabled on the
command-line.
- (nm_device_update_link_active): split out logic for wired and wireless device link
checking to separate functions to facilitate test device link checking.
- (nm_device_set_enc_key): Since some drivers for wireless cards are daft and
don't make a distinction between System Authentication and Encryption
(namely Cisco aironet), we use Open System auth when setting a WEP key
on the card. We don't deal with Shared Key auth yet.
- (nm_device_activation_worker): split the activation cancel check logic out into
a separate routine nm_device_activation_cancel_if_needed()
- (nm_device_activation_signal_cancel): rename from nm_device_activation_cancel()
- (nm_device_fake_ap_list): Test wireless devices obviously cannot scan, so create
a list of fake access points that they can "see"
- (nm_device_is_test_device): return whether or not a device is a test device
* src/NetworkManagerPolicy.c
- (nm_policy_get_best_device): attempt to deal with wireless network selection,
previously if you "locked"/forced NM to use a wireless device but then
selected a wireless network for NM to use, it would switch to a wired device.
So, if the active device is wireless and it has a "forced" best AP, use it
if the "forced" best AP is still valid
- (nm_state_modification_monitor): deal with NULL best devices, for example
there were no usable network devices, or the last one was removed
* src/backends/NetworkManager*.c
- Deal with test devices, mostly just return success for operations like getting
a DHCP address
* test/nmtestdevices.c
- Test tool to create/remove/link-switch test devices
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@112 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-31 16:09:15 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
2004-09-08 18:14:42 +00:00
|
|
|
/*
|
|
|
|
|
* nm_device_activation_configure_ip
|
|
|
|
|
*
|
|
|
|
|
* Perform any IP-based configuration on a device, like running DHCP
|
|
|
|
|
* or manually setting up the IP address, gateway, and default route.
|
|
|
|
|
*
|
|
|
|
|
*/
|
2004-10-08 05:22:17 +00:00
|
|
|
static gboolean nm_device_activation_configure_ip (NMDevice *dev)
|
2004-09-08 18:14:42 +00:00
|
|
|
{
|
2004-10-08 05:22:17 +00:00
|
|
|
gboolean success = FALSE;
|
|
|
|
|
|
|
|
|
|
g_return_val_if_fail (dev != NULL, FALSE);
|
2004-09-08 18:14:42 +00:00
|
|
|
|
|
|
|
|
if (nm_device_config_get_use_dhcp (dev))
|
|
|
|
|
{
|
2004-10-08 05:22:17 +00:00
|
|
|
if (nm_system_device_run_dhcp (dev))
|
|
|
|
|
success = TRUE;
|
|
|
|
|
else
|
2004-09-08 18:14:42 +00:00
|
|
|
{
|
|
|
|
|
/* Interfaces cannot be down if they are the active interface,
|
|
|
|
|
* otherwise we cannot use them for scanning or link detection.
|
|
|
|
|
* If dhclient doesn't get a DHCP address, it will take the interface
|
|
|
|
|
* down, so we reactivate it here.
|
|
|
|
|
*/
|
|
|
|
|
if (nm_device_is_wireless (dev))
|
|
|
|
|
{
|
|
|
|
|
nm_device_set_essid (dev, "");
|
2004-11-03 07:47:32 +00:00
|
|
|
nm_device_set_enc_key (dev, NULL, NM_DEVICE_AUTH_METHOD_NONE);
|
2004-09-08 18:14:42 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
nm_device_bring_up (dev);
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
else
|
|
|
|
|
{
|
|
|
|
|
/* Manually set up the device */
|
2004-10-12 11:15:47 +00:00
|
|
|
success = nm_system_device_setup_static_ip4_config (dev);
|
2004-09-08 18:14:42 +00:00
|
|
|
}
|
2004-10-08 05:22:17 +00:00
|
|
|
|
|
|
|
|
return (success);
|
2004-09-08 18:14:42 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
/*
|
|
|
|
|
* nm_device_activation_worker
|
|
|
|
|
*
|
|
|
|
|
* Thread worker function to actually activate a device. We have to do it in another
|
|
|
|
|
* thread because things like dhclient block our main thread's event loop, and thus we
|
|
|
|
|
* wouldn't respond to dbus messages.
|
|
|
|
|
*/
|
|
|
|
|
static gpointer nm_device_activation_worker (gpointer user_data)
|
|
|
|
|
{
|
2004-08-12 Dan Williams <dcbw@redhat.com>
* info-daemon/passphrase.glade
- Set window title to " "
* panel-applet/Makefile.am
panel-applet/keyring.png
- Deliver to correct place
* panel-applet/NMWirelessApplet.[ch]
- Add comments
- Remove applet->have_active_device as its no longer used
- (nmwa_load_theme): load keyring.png too
- (error_dialog): remove
- (show_warning_dialog): subsume functionality of error dialog too
- (nmwa_destroy, nmwa_new): create and dispose of an application-wide GConfClient
- (nmwa_handle_network_choice): add to deal with user clicking on an item from
the networks menu
- (nmwa_menu_item_activated): GtkMenuItem "activate" signal handler
- (nmwa_button_clicked, nmwa_setup_widgets): create and populate the menu on startup
and when we get broadcasts of changed wireless access points only, not when the
user clicks on the button to display the menu (too long of a wait)
- (nmwa_add_menu_item): Make active network bold, and place a keyring icon beside
networks that are encrypted
- (nmwa_dispose_menu, nmwa_menu_item_data_free): dispose of the data we place on each
menu item with g_object_set_data()
* panel-applet/NMWirelessAppletDbus.[ch]
- (nmwa_dbus_get_bool): add method to return boolean value from dbus message
- (nmwa_dbus_get_active_network): add (nmwa_dbus_get_string() wrapper to get active network)
- (nmwa_dbus_add_networks_to_menu): clean up, only show one instance of each ESSID in the menu
- (nmwa_dbus_set_network): force NetworkManager to use a particular network for wireless cards
- (nmwa_dbus_init, nmwa_dbus_filter): Trap network appear/disappear and device
activation/deactivation signals and rebuild the menu when they happen
* src/NetworkManager.c
- (main): use new nm_spawn_process() rather than system()
* src/NetworkManagerDbus.c
- (nm_dbus_devices_handle_request): don't compare AP structure addresses directly, but essids
instead. Since we can now force best_aps to stick around, the AP structure to which
dev->options.wireless.best_ap points to won't necessarily be in the device's device list
if a scan has happened since the best_ap was frozen. Also add "setNetwork" method
to freeze the best_ap.
* src/NetworkManagerDevice.[ch]
- (nm_device_activation_worker): Use new nm_spawn_process() call rather than system()
- (nm_device_*_best_ap): add freeze/unfreeze/get_frozen functions, and don't really update
the best_ap in nm_device_update_best_ap() if the best_ap is frozen AND in the device's
ap list
* src/NetworkManagerUtils.[ch]
- (nm_spawn_process): add replacement for system() usage
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@48 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-12 19:58:01 +00:00
|
|
|
NMDevice *dev = (NMDevice *)user_data;
|
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
|
|
|
unsigned char hostname[100] = "\0";
|
2004-07-06 01:34:10 +00:00
|
|
|
int host_err;
|
2004-10-29 16:37:43 +00:00
|
|
|
NMAccessPoint *best_ap = NULL;
|
2004-07-06 01:34:10 +00:00
|
|
|
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
g_return_val_if_fail (dev != NULL, NULL);
|
|
|
|
|
g_return_val_if_fail (dev->app_data != NULL, NULL);
|
|
|
|
|
|
2004-08-24 00:31:47 +00:00
|
|
|
syslog (LOG_DEBUG, "nm_device_activation_worker (%s) started...", nm_device_get_iface (dev));
|
2004-07-06 01:34:10 +00:00
|
|
|
|
|
|
|
|
/* If its a wireless device, set the ESSID and WEP key */
|
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
|
|
|
if (nm_device_is_wireless (dev))
|
2004-07-06 01:34:10 +00:00
|
|
|
{
|
2004-09-09 20:02:59 +00:00
|
|
|
nm_device_activate_wireless_wait_for_link (dev);
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
|
2004-09-09 20:02:59 +00:00
|
|
|
/* If we were told to quit activation, stop the thread and return */
|
2004-09-12 03:06:44 +00:00
|
|
|
if (nm_device_activation_should_cancel (dev))
|
2004-10-23 06:43:01 +00:00
|
|
|
goto out;
|
2004-08-24 00:31:47 +00:00
|
|
|
|
2004-10-29 16:37:43 +00:00
|
|
|
best_ap = nm_device_get_best_ap (dev);
|
2004-08-24 00:31:47 +00:00
|
|
|
syslog (LOG_DEBUG, "nm_device_activation_worker(%s): using ESSID '%s'", nm_device_get_iface (dev),
|
2004-10-29 16:37:43 +00:00
|
|
|
best_ap ? nm_ap_get_essid (best_ap) : "none");
|
|
|
|
|
if (best_ap)
|
|
|
|
|
nm_ap_unref (best_ap);
|
2004-07-06 01:34:10 +00:00
|
|
|
}
|
2004-07-19 Dan Williams <dcbw@redhat.com>
* Makefile.am
- Add info-daemon directory
* configure.in
- Check for glade libs and headers
- Add info-daemon directory
* src/NetworkManagerAP.c
- nm_ap_new_from_ap(): Fix bug that resulted in an APs encryption status not getting
copied over to the new AP.
* src/NetworkManagerDbus.c
src/NetworkManagerDbus.h
- Deal with nm_device_ap_list_get_ap()->nm_device_ap_list_get_ap_by_index() change
- Remove nm_dbus_signal_need_key_for_network()
- Add disabled code for asynchronous user wep key callbacks
- Add functions for getting, setting, and cancelling user key operations
- Remove "setKeyForNetwork" device dbus method call, its on NetworkManager object instead
- Add "setKeyForNetwork" dbus method call on NetworkManager object
* src/NetworkManagerDevice.c
src/NetworkManagerDevice.h
- nm_device_update_link_active(): revert changes for wireless link detection, the WEP-key-is-wrong
logic is in device activation now
- nm_device_activate(): for wireless devices, if we can't associate with access point (perhaps
key is wrong) trigger get-user-key pending action
- Implement get-user-key pending action stuff, tie to dbus messages
- Rename nm_device_ap_list_get_ap() -> nm_device_ap_list_get_ap_by_index()
- Add nm_device_ap_list_get_ap_by_essid()
- Instead of copying "best" access points, ref them instead so that the key we set
sticks around
* src/NetworkManagerPolicy.c
- Deal with wrong WEP key, but right access point (and if so, return link_active = TRUE)
- Don't cancel pending actions on a device if its the same device as last iteration
- Only promote pending_device->active_device if activation was successfull
* src/Makefile.am
- Rename nmclienttest->nmtest
* info-daemon/Makefile.am
info-daemon/NetworkManagerInfo.c
info-daemon/NetworkManagerInfo.h
info-daemon/NetworkManagerInfoDbus.c
info-daemon/NetworkManagerInfoDbus.h
info-daemon/passphrase.glade
info-daemon/NetworkManagerInfo.conf
info-daemon/keyring.png
- Import sources for info-daemon, which pops up dialog for passphrase/key when
NetworkManager asks for it, and also will (soon) provide "allowed" access point
lists to NetworkManager by proxying user's GConf
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@16 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-07-19 06:08:52 +00:00
|
|
|
else
|
|
|
|
|
{
|
|
|
|
|
/* Bring the device up */
|
|
|
|
|
if (!nm_device_is_up (dev));
|
|
|
|
|
nm_device_bring_up (dev);
|
|
|
|
|
}
|
2004-07-06 01:34:10 +00:00
|
|
|
|
2004-08-24 00:31:47 +00:00
|
|
|
nm_system_delete_default_route ();
|
|
|
|
|
nm_system_device_stop_dhcp (dev);
|
2004-07-06 01:34:10 +00:00
|
|
|
|
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
|
|
|
/* If we don't have a "best" ap, don't try to get a DHCP address or restart the name service cache */
|
2004-10-29 16:37:43 +00:00
|
|
|
if (nm_device_is_wireless (dev))
|
|
|
|
|
best_ap = nm_device_get_best_ap (dev);
|
|
|
|
|
if (nm_device_is_wired (dev) || best_ap)
|
2004-07-06 01:34:10 +00:00
|
|
|
{
|
2004-10-08 05:22:17 +00:00
|
|
|
gboolean success;
|
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
|
|
|
/* Save machine host name */
|
|
|
|
|
host_err = gethostname (&hostname[0], 100);
|
|
|
|
|
|
2004-10-08 05:22:17 +00:00
|
|
|
if (!(success = nm_device_activation_configure_ip (dev)))
|
|
|
|
|
syslog (LOG_DEBUG, "nm_device_activation_worker(%s): could not retrieve and assign IP information to device\n", nm_device_get_iface (dev));
|
2004-07-06 01:34:10 +00:00
|
|
|
|
2004-08-24 00:31:47 +00:00
|
|
|
/* Set the hostname back to what it was before so that X11 doesn't
|
|
|
|
|
* puke when the hostname changes, and so users can actually launch stuff.
|
|
|
|
|
*/
|
|
|
|
|
if (host_err >= 0)
|
|
|
|
|
sethostname (hostname, strlen (hostname));
|
|
|
|
|
|
2004-08-12 Dan Williams <dcbw@redhat.com>
* info-daemon/passphrase.glade
- Set window title to " "
* panel-applet/Makefile.am
panel-applet/keyring.png
- Deliver to correct place
* panel-applet/NMWirelessApplet.[ch]
- Add comments
- Remove applet->have_active_device as its no longer used
- (nmwa_load_theme): load keyring.png too
- (error_dialog): remove
- (show_warning_dialog): subsume functionality of error dialog too
- (nmwa_destroy, nmwa_new): create and dispose of an application-wide GConfClient
- (nmwa_handle_network_choice): add to deal with user clicking on an item from
the networks menu
- (nmwa_menu_item_activated): GtkMenuItem "activate" signal handler
- (nmwa_button_clicked, nmwa_setup_widgets): create and populate the menu on startup
and when we get broadcasts of changed wireless access points only, not when the
user clicks on the button to display the menu (too long of a wait)
- (nmwa_add_menu_item): Make active network bold, and place a keyring icon beside
networks that are encrypted
- (nmwa_dispose_menu, nmwa_menu_item_data_free): dispose of the data we place on each
menu item with g_object_set_data()
* panel-applet/NMWirelessAppletDbus.[ch]
- (nmwa_dbus_get_bool): add method to return boolean value from dbus message
- (nmwa_dbus_get_active_network): add (nmwa_dbus_get_string() wrapper to get active network)
- (nmwa_dbus_add_networks_to_menu): clean up, only show one instance of each ESSID in the menu
- (nmwa_dbus_set_network): force NetworkManager to use a particular network for wireless cards
- (nmwa_dbus_init, nmwa_dbus_filter): Trap network appear/disappear and device
activation/deactivation signals and rebuild the menu when they happen
* src/NetworkManager.c
- (main): use new nm_spawn_process() rather than system()
* src/NetworkManagerDbus.c
- (nm_dbus_devices_handle_request): don't compare AP structure addresses directly, but essids
instead. Since we can now force best_aps to stick around, the AP structure to which
dev->options.wireless.best_ap points to won't necessarily be in the device's device list
if a scan has happened since the best_ap was frozen. Also add "setNetwork" method
to freeze the best_ap.
* src/NetworkManagerDevice.[ch]
- (nm_device_activation_worker): Use new nm_spawn_process() call rather than system()
- (nm_device_*_best_ap): add freeze/unfreeze/get_frozen functions, and don't really update
the best_ap in nm_device_update_best_ap() if the best_ap is frozen AND in the device's
ap list
* src/NetworkManagerUtils.[ch]
- (nm_spawn_process): add replacement for system() usage
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@48 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-12 19:58:01 +00:00
|
|
|
/* If we were told to quit activation, stop the thread and return */
|
2004-10-23 06:43:01 +00:00
|
|
|
if (nm_device_activation_should_cancel (dev))
|
|
|
|
|
goto out;
|
|
|
|
|
|
|
|
|
|
if (!success)
|
2004-09-28 19:50:37 +00:00
|
|
|
{
|
2004-10-23 06:43:01 +00:00
|
|
|
dev->activating = FALSE;
|
|
|
|
|
dev->just_activated = FALSE;
|
|
|
|
|
dev->activation_failed = TRUE;
|
|
|
|
|
dev->quit_activation = FALSE;
|
|
|
|
|
goto out;
|
2004-09-28 19:50:37 +00:00
|
|
|
}
|
2004-08-12 Dan Williams <dcbw@redhat.com>
* info-daemon/passphrase.glade
- Set window title to " "
* panel-applet/Makefile.am
panel-applet/keyring.png
- Deliver to correct place
* panel-applet/NMWirelessApplet.[ch]
- Add comments
- Remove applet->have_active_device as its no longer used
- (nmwa_load_theme): load keyring.png too
- (error_dialog): remove
- (show_warning_dialog): subsume functionality of error dialog too
- (nmwa_destroy, nmwa_new): create and dispose of an application-wide GConfClient
- (nmwa_handle_network_choice): add to deal with user clicking on an item from
the networks menu
- (nmwa_menu_item_activated): GtkMenuItem "activate" signal handler
- (nmwa_button_clicked, nmwa_setup_widgets): create and populate the menu on startup
and when we get broadcasts of changed wireless access points only, not when the
user clicks on the button to display the menu (too long of a wait)
- (nmwa_add_menu_item): Make active network bold, and place a keyring icon beside
networks that are encrypted
- (nmwa_dispose_menu, nmwa_menu_item_data_free): dispose of the data we place on each
menu item with g_object_set_data()
* panel-applet/NMWirelessAppletDbus.[ch]
- (nmwa_dbus_get_bool): add method to return boolean value from dbus message
- (nmwa_dbus_get_active_network): add (nmwa_dbus_get_string() wrapper to get active network)
- (nmwa_dbus_add_networks_to_menu): clean up, only show one instance of each ESSID in the menu
- (nmwa_dbus_set_network): force NetworkManager to use a particular network for wireless cards
- (nmwa_dbus_init, nmwa_dbus_filter): Trap network appear/disappear and device
activation/deactivation signals and rebuild the menu when they happen
* src/NetworkManager.c
- (main): use new nm_spawn_process() rather than system()
* src/NetworkManagerDbus.c
- (nm_dbus_devices_handle_request): don't compare AP structure addresses directly, but essids
instead. Since we can now force best_aps to stick around, the AP structure to which
dev->options.wireless.best_ap points to won't necessarily be in the device's device list
if a scan has happened since the best_ap was frozen. Also add "setNetwork" method
to freeze the best_ap.
* src/NetworkManagerDevice.[ch]
- (nm_device_activation_worker): Use new nm_spawn_process() call rather than system()
- (nm_device_*_best_ap): add freeze/unfreeze/get_frozen functions, and don't really update
the best_ap in nm_device_update_best_ap() if the best_ap is frozen AND in the device's
ap list
* src/NetworkManagerUtils.[ch]
- (nm_spawn_process): add replacement for system() usage
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@48 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-12 19:58:01 +00:00
|
|
|
|
2004-08-24 00:31:47 +00:00
|
|
|
/* Make system aware of any new DNS settings from resolv.conf */
|
|
|
|
|
nm_system_update_dns ();
|
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
|
|
|
}
|
2004-10-29 16:37:43 +00:00
|
|
|
if (best_ap)
|
|
|
|
|
nm_ap_unref (best_ap);
|
2004-07-06 01:34:10 +00:00
|
|
|
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
/* If we were told to quit activation, stop the thread and return */
|
2004-09-12 03:06:44 +00:00
|
|
|
if (nm_device_activation_should_cancel (dev))
|
2004-10-23 06:43:01 +00:00
|
|
|
goto out;
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
|
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
|
|
|
nm_device_update_ip4_address (dev);
|
2004-07-06 01:34:10 +00:00
|
|
|
|
2004-10-23 06:43:01 +00:00
|
|
|
dev->just_activated = TRUE;
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
dev->activating = FALSE;
|
2004-10-23 06:43:01 +00:00
|
|
|
dev->activation_failed = FALSE;
|
|
|
|
|
dev->quit_activation = FALSE;
|
|
|
|
|
|
|
|
|
|
syslog (LOG_DEBUG, "nm_device_activation_worker(%s): device activated", nm_device_get_iface (dev));
|
|
|
|
|
|
|
|
|
|
out:
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
nm_device_unref (dev);
|
|
|
|
|
return (NULL);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
2004-10-11 21:32:19 +00:00
|
|
|
* nm_device_is_just_activated
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
*
|
|
|
|
|
* Check if the device was just activated successfully or not. If so, clear
|
|
|
|
|
* its just_activated flag and return TRUE. If its not activated yet, return FALSE.
|
|
|
|
|
*
|
|
|
|
|
*/
|
2004-10-11 21:32:19 +00:00
|
|
|
gboolean nm_device_is_just_activated (NMDevice *dev)
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
{
|
|
|
|
|
g_return_val_if_fail (dev != NULL, FALSE);
|
|
|
|
|
|
|
|
|
|
if (dev->just_activated)
|
|
|
|
|
{
|
|
|
|
|
dev->just_activated = FALSE;
|
|
|
|
|
return (TRUE);
|
|
|
|
|
}
|
2004-08-31 Dan Williams <dcbw@redhat.com>
* Remove 'debug' extern global from all files since we now
use syslog()
* src/NetworkManager.[ch]
- Break out routine that get the net.interface property from HAL,
removing that logic from nm_create_device_and_add_to_list()
- (nm_create_device_and_add_to_list): make this a bit more general so
it doesn't do the talking to HAL. Also add arguments to facilitate
the create of test devices.
- (nm_data_mark_state_changed): rename from nm_data_set_state_modified()
- (nm_data_new, main, nm_print_usage): add new argument "--enable-test-devices"
which makes NetworkManager listen for dbus commands to create test
devices, which have no backing hardware. Use when you're on a plane
for example, and/or forgot your wireless card at home. Test devices
_cannot_ be created unless NM is started with --enable-test-devices.
* src/NetworkManagerDbus.[ch]
- New "getLinkActive" method for devices
- New "setLinkActive" method for devices (only works on test devices)
- New "createTestDevice" method on NetworkManager object to create a test
device of a specified type (ie wired, wireless). UDI is created from
scratch, as is the interface name. Only works when NM is started with
--enable-test-devices switch.
- New "removeTestDevice" method on NetworkManager object which removes a
test device. Only works when NM is started with --enable-test-devices
* src/NetworkManagerDevice.[ch]
- Logic to facilitate test devices. Add variables to NMDevice struct to indicate
whether a device is a test device or not, and what its link status is.
- Deal with test devices in most functions. For those that work directly on hardware
special-case test devices.
- (nm_device_new): don't create a test device if test devices weren't enabled on the
command-line.
- (nm_device_update_link_active): split out logic for wired and wireless device link
checking to separate functions to facilitate test device link checking.
- (nm_device_set_enc_key): Since some drivers for wireless cards are daft and
don't make a distinction between System Authentication and Encryption
(namely Cisco aironet), we use Open System auth when setting a WEP key
on the card. We don't deal with Shared Key auth yet.
- (nm_device_activation_worker): split the activation cancel check logic out into
a separate routine nm_device_activation_cancel_if_needed()
- (nm_device_activation_signal_cancel): rename from nm_device_activation_cancel()
- (nm_device_fake_ap_list): Test wireless devices obviously cannot scan, so create
a list of fake access points that they can "see"
- (nm_device_is_test_device): return whether or not a device is a test device
* src/NetworkManagerPolicy.c
- (nm_policy_get_best_device): attempt to deal with wireless network selection,
previously if you "locked"/forced NM to use a wireless device but then
selected a wireless network for NM to use, it would switch to a wired device.
So, if the active device is wireless and it has a "forced" best AP, use it
if the "forced" best AP is still valid
- (nm_state_modification_monitor): deal with NULL best devices, for example
there were no usable network devices, or the last one was removed
* src/backends/NetworkManager*.c
- Deal with test devices, mostly just return success for operations like getting
a DHCP address
* test/nmtestdevices.c
- Test tool to create/remove/link-switch test devices
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@112 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-31 16:09:15 +00:00
|
|
|
|
|
|
|
|
return (FALSE);
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
2004-10-11 21:32:19 +00:00
|
|
|
* nm_device_is_activating
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
*
|
|
|
|
|
* Return whether or not the device is currently activating itself.
|
|
|
|
|
*
|
|
|
|
|
*/
|
2004-10-11 21:32:19 +00:00
|
|
|
gboolean nm_device_is_activating (NMDevice *dev)
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
{
|
|
|
|
|
g_return_val_if_fail (dev != NULL, FALSE);
|
|
|
|
|
|
|
|
|
|
return (dev->activating);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
2004-10-23 06:43:01 +00:00
|
|
|
/*
|
|
|
|
|
* nm_device_did_activation_fail
|
|
|
|
|
*
|
|
|
|
|
* Return whether or not activation on the device failed (was cancelled or
|
|
|
|
|
* activation process encountered and error)
|
|
|
|
|
*
|
|
|
|
|
*/
|
|
|
|
|
gboolean nm_device_did_activation_fail (NMDevice *dev)
|
|
|
|
|
{
|
|
|
|
|
g_return_val_if_fail (dev != NULL, FALSE);
|
|
|
|
|
|
|
|
|
|
return (dev->activation_failed);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
/*
|
2004-09-12 03:06:44 +00:00
|
|
|
* nm_device_activation_cancel
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
*
|
|
|
|
|
* Signal activation worker that it should stop and die.
|
|
|
|
|
*
|
|
|
|
|
*/
|
2004-09-12 03:06:44 +00:00
|
|
|
void nm_device_activation_cancel (NMDevice *dev)
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
{
|
|
|
|
|
g_return_if_fail (dev != NULL);
|
|
|
|
|
|
2004-10-11 21:32:19 +00:00
|
|
|
if (nm_device_is_activating (dev))
|
2004-08-31 Dan Williams <dcbw@redhat.com>
* Remove 'debug' extern global from all files since we now
use syslog()
* src/NetworkManager.[ch]
- Break out routine that get the net.interface property from HAL,
removing that logic from nm_create_device_and_add_to_list()
- (nm_create_device_and_add_to_list): make this a bit more general so
it doesn't do the talking to HAL. Also add arguments to facilitate
the create of test devices.
- (nm_data_mark_state_changed): rename from nm_data_set_state_modified()
- (nm_data_new, main, nm_print_usage): add new argument "--enable-test-devices"
which makes NetworkManager listen for dbus commands to create test
devices, which have no backing hardware. Use when you're on a plane
for example, and/or forgot your wireless card at home. Test devices
_cannot_ be created unless NM is started with --enable-test-devices.
* src/NetworkManagerDbus.[ch]
- New "getLinkActive" method for devices
- New "setLinkActive" method for devices (only works on test devices)
- New "createTestDevice" method on NetworkManager object to create a test
device of a specified type (ie wired, wireless). UDI is created from
scratch, as is the interface name. Only works when NM is started with
--enable-test-devices switch.
- New "removeTestDevice" method on NetworkManager object which removes a
test device. Only works when NM is started with --enable-test-devices
* src/NetworkManagerDevice.[ch]
- Logic to facilitate test devices. Add variables to NMDevice struct to indicate
whether a device is a test device or not, and what its link status is.
- Deal with test devices in most functions. For those that work directly on hardware
special-case test devices.
- (nm_device_new): don't create a test device if test devices weren't enabled on the
command-line.
- (nm_device_update_link_active): split out logic for wired and wireless device link
checking to separate functions to facilitate test device link checking.
- (nm_device_set_enc_key): Since some drivers for wireless cards are daft and
don't make a distinction between System Authentication and Encryption
(namely Cisco aironet), we use Open System auth when setting a WEP key
on the card. We don't deal with Shared Key auth yet.
- (nm_device_activation_worker): split the activation cancel check logic out into
a separate routine nm_device_activation_cancel_if_needed()
- (nm_device_activation_signal_cancel): rename from nm_device_activation_cancel()
- (nm_device_fake_ap_list): Test wireless devices obviously cannot scan, so create
a list of fake access points that they can "see"
- (nm_device_is_test_device): return whether or not a device is a test device
* src/NetworkManagerPolicy.c
- (nm_policy_get_best_device): attempt to deal with wireless network selection,
previously if you "locked"/forced NM to use a wireless device but then
selected a wireless network for NM to use, it would switch to a wired device.
So, if the active device is wireless and it has a "forced" best AP, use it
if the "forced" best AP is still valid
- (nm_state_modification_monitor): deal with NULL best devices, for example
there were no usable network devices, or the last one was removed
* src/backends/NetworkManager*.c
- Deal with test devices, mostly just return success for operations like getting
a DHCP address
* test/nmtestdevices.c
- Test tool to create/remove/link-switch test devices
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@112 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-31 16:09:15 +00:00
|
|
|
{
|
2004-09-12 03:06:44 +00:00
|
|
|
syslog (LOG_DEBUG, "nm_device_activation_cancel(%s): cancelling...", nm_device_get_iface (dev));
|
2004-08-31 Dan Williams <dcbw@redhat.com>
* Remove 'debug' extern global from all files since we now
use syslog()
* src/NetworkManager.[ch]
- Break out routine that get the net.interface property from HAL,
removing that logic from nm_create_device_and_add_to_list()
- (nm_create_device_and_add_to_list): make this a bit more general so
it doesn't do the talking to HAL. Also add arguments to facilitate
the create of test devices.
- (nm_data_mark_state_changed): rename from nm_data_set_state_modified()
- (nm_data_new, main, nm_print_usage): add new argument "--enable-test-devices"
which makes NetworkManager listen for dbus commands to create test
devices, which have no backing hardware. Use when you're on a plane
for example, and/or forgot your wireless card at home. Test devices
_cannot_ be created unless NM is started with --enable-test-devices.
* src/NetworkManagerDbus.[ch]
- New "getLinkActive" method for devices
- New "setLinkActive" method for devices (only works on test devices)
- New "createTestDevice" method on NetworkManager object to create a test
device of a specified type (ie wired, wireless). UDI is created from
scratch, as is the interface name. Only works when NM is started with
--enable-test-devices switch.
- New "removeTestDevice" method on NetworkManager object which removes a
test device. Only works when NM is started with --enable-test-devices
* src/NetworkManagerDevice.[ch]
- Logic to facilitate test devices. Add variables to NMDevice struct to indicate
whether a device is a test device or not, and what its link status is.
- Deal with test devices in most functions. For those that work directly on hardware
special-case test devices.
- (nm_device_new): don't create a test device if test devices weren't enabled on the
command-line.
- (nm_device_update_link_active): split out logic for wired and wireless device link
checking to separate functions to facilitate test device link checking.
- (nm_device_set_enc_key): Since some drivers for wireless cards are daft and
don't make a distinction between System Authentication and Encryption
(namely Cisco aironet), we use Open System auth when setting a WEP key
on the card. We don't deal with Shared Key auth yet.
- (nm_device_activation_worker): split the activation cancel check logic out into
a separate routine nm_device_activation_cancel_if_needed()
- (nm_device_activation_signal_cancel): rename from nm_device_activation_cancel()
- (nm_device_fake_ap_list): Test wireless devices obviously cannot scan, so create
a list of fake access points that they can "see"
- (nm_device_is_test_device): return whether or not a device is a test device
* src/NetworkManagerPolicy.c
- (nm_policy_get_best_device): attempt to deal with wireless network selection,
previously if you "locked"/forced NM to use a wireless device but then
selected a wireless network for NM to use, it would switch to a wired device.
So, if the active device is wireless and it has a "forced" best AP, use it
if the "forced" best AP is still valid
- (nm_state_modification_monitor): deal with NULL best devices, for example
there were no usable network devices, or the last one was removed
* src/backends/NetworkManager*.c
- Deal with test devices, mostly just return success for operations like getting
a DHCP address
* test/nmtestdevices.c
- Test tool to create/remove/link-switch test devices
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@112 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-31 16:09:15 +00:00
|
|
|
dev->quit_activation = TRUE;
|
2004-09-09 21:34:40 +00:00
|
|
|
nm_system_kill_all_dhcp_daemons (); /* dhcp daemons will block, so have to kill them to return control */
|
2004-09-12 03:06:44 +00:00
|
|
|
|
|
|
|
|
/* Spin until cancelled. Possible race conditions or deadlocks here.
|
|
|
|
|
* The other problem with waiting here is that we hold up dbus traffic
|
|
|
|
|
* that we should respond to.
|
|
|
|
|
*/
|
2004-10-11 21:32:19 +00:00
|
|
|
while (nm_device_is_activating (dev))
|
2004-09-12 03:06:44 +00:00
|
|
|
g_usleep (G_USEC_PER_SEC / 2);
|
2004-08-31 Dan Williams <dcbw@redhat.com>
* Remove 'debug' extern global from all files since we now
use syslog()
* src/NetworkManager.[ch]
- Break out routine that get the net.interface property from HAL,
removing that logic from nm_create_device_and_add_to_list()
- (nm_create_device_and_add_to_list): make this a bit more general so
it doesn't do the talking to HAL. Also add arguments to facilitate
the create of test devices.
- (nm_data_mark_state_changed): rename from nm_data_set_state_modified()
- (nm_data_new, main, nm_print_usage): add new argument "--enable-test-devices"
which makes NetworkManager listen for dbus commands to create test
devices, which have no backing hardware. Use when you're on a plane
for example, and/or forgot your wireless card at home. Test devices
_cannot_ be created unless NM is started with --enable-test-devices.
* src/NetworkManagerDbus.[ch]
- New "getLinkActive" method for devices
- New "setLinkActive" method for devices (only works on test devices)
- New "createTestDevice" method on NetworkManager object to create a test
device of a specified type (ie wired, wireless). UDI is created from
scratch, as is the interface name. Only works when NM is started with
--enable-test-devices switch.
- New "removeTestDevice" method on NetworkManager object which removes a
test device. Only works when NM is started with --enable-test-devices
* src/NetworkManagerDevice.[ch]
- Logic to facilitate test devices. Add variables to NMDevice struct to indicate
whether a device is a test device or not, and what its link status is.
- Deal with test devices in most functions. For those that work directly on hardware
special-case test devices.
- (nm_device_new): don't create a test device if test devices weren't enabled on the
command-line.
- (nm_device_update_link_active): split out logic for wired and wireless device link
checking to separate functions to facilitate test device link checking.
- (nm_device_set_enc_key): Since some drivers for wireless cards are daft and
don't make a distinction between System Authentication and Encryption
(namely Cisco aironet), we use Open System auth when setting a WEP key
on the card. We don't deal with Shared Key auth yet.
- (nm_device_activation_worker): split the activation cancel check logic out into
a separate routine nm_device_activation_cancel_if_needed()
- (nm_device_activation_signal_cancel): rename from nm_device_activation_cancel()
- (nm_device_fake_ap_list): Test wireless devices obviously cannot scan, so create
a list of fake access points that they can "see"
- (nm_device_is_test_device): return whether or not a device is a test device
* src/NetworkManagerPolicy.c
- (nm_policy_get_best_device): attempt to deal with wireless network selection,
previously if you "locked"/forced NM to use a wireless device but then
selected a wireless network for NM to use, it would switch to a wired device.
So, if the active device is wireless and it has a "forced" best AP, use it
if the "forced" best AP is still valid
- (nm_state_modification_monitor): deal with NULL best devices, for example
there were no usable network devices, or the last one was removed
* src/backends/NetworkManager*.c
- Deal with test devices, mostly just return success for operations like getting
a DHCP address
* test/nmtestdevices.c
- Test tool to create/remove/link-switch test devices
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@112 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-31 16:09:15 +00:00
|
|
|
}
|
2004-07-06 01:34:10 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* nm_device_deactivate
|
|
|
|
|
*
|
|
|
|
|
* Remove a device's routing table entries and IP address.
|
|
|
|
|
*
|
|
|
|
|
*/
|
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
|
|
|
gboolean nm_device_deactivate (NMDevice *dev, gboolean just_added)
|
2004-07-06 01:34:10 +00:00
|
|
|
{
|
|
|
|
|
g_return_val_if_fail (dev != NULL, FALSE);
|
2004-07-25 02:40:19 +00:00
|
|
|
g_return_val_if_fail (dev->app_data != NULL, FALSE);
|
2004-07-06 01:34:10 +00:00
|
|
|
|
2004-09-12 03:06:44 +00:00
|
|
|
nm_device_activation_cancel (dev);
|
2004-07-06 01:34:10 +00:00
|
|
|
|
2004-10-21 19:00:13 +00:00
|
|
|
if (nm_device_get_driver_support_level (dev) == NM_DRIVER_UNSUPPORTED)
|
|
|
|
|
return (TRUE);
|
|
|
|
|
|
2004-08-24 00:31:47 +00:00
|
|
|
/* Take out any entries in the routing table and any IP address the old device had. */
|
|
|
|
|
nm_system_device_flush_routes (dev);
|
|
|
|
|
nm_system_device_flush_addresses (dev);
|
|
|
|
|
dev->ip4_address = 0;
|
2004-07-06 01:34:10 +00:00
|
|
|
|
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
|
|
|
if (!just_added)
|
2004-08-26 20:05:24 +00:00
|
|
|
nm_dbus_signal_device_status_change (dev->app_data->dbus_connection, dev, DEVICE_NO_LONGER_ACTIVE);
|
2004-07-06 01:34:10 +00:00
|
|
|
|
2004-09-09 20:02:59 +00:00
|
|
|
/* Clean up stuff, don't leave the card associated */
|
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
|
|
|
if (nm_device_is_wireless (dev))
|
2004-07-06 04:45:00 +00:00
|
|
|
{
|
|
|
|
|
nm_device_set_essid (dev, "");
|
2004-11-03 07:47:32 +00:00
|
|
|
nm_device_set_enc_key (dev, NULL, NM_DEVICE_AUTH_METHOD_NONE);
|
2004-07-06 04:45:00 +00:00
|
|
|
}
|
|
|
|
|
|
2004-08-24 00:31:47 +00:00
|
|
|
return (TRUE);
|
2004-07-06 01:34:10 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
2004-10-07 01:50:39 +00:00
|
|
|
/*
|
2004-10-11 21:32:19 +00:00
|
|
|
* nm_device_is_scanning
|
2004-10-07 01:50:39 +00:00
|
|
|
*
|
|
|
|
|
* Returns whether the device is scanning, awaiting an access point to connect to.
|
|
|
|
|
* Note that this does NOT get set when the device is actually scanning, just
|
|
|
|
|
* when it is waiting for a valid access point to connect to.
|
|
|
|
|
*
|
|
|
|
|
*/
|
2004-10-11 21:32:19 +00:00
|
|
|
gboolean nm_device_is_scanning (NMDevice *dev)
|
2004-10-07 01:50:39 +00:00
|
|
|
{
|
|
|
|
|
g_return_val_if_fail (dev != NULL, FALSE);
|
|
|
|
|
g_return_val_if_fail (nm_device_is_wireless (dev), FALSE);
|
|
|
|
|
|
|
|
|
|
return (dev->options.wireless.now_scanning);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
2004-07-19 Dan Williams <dcbw@redhat.com>
* Makefile.am
- Add info-daemon directory
* configure.in
- Check for glade libs and headers
- Add info-daemon directory
* src/NetworkManagerAP.c
- nm_ap_new_from_ap(): Fix bug that resulted in an APs encryption status not getting
copied over to the new AP.
* src/NetworkManagerDbus.c
src/NetworkManagerDbus.h
- Deal with nm_device_ap_list_get_ap()->nm_device_ap_list_get_ap_by_index() change
- Remove nm_dbus_signal_need_key_for_network()
- Add disabled code for asynchronous user wep key callbacks
- Add functions for getting, setting, and cancelling user key operations
- Remove "setKeyForNetwork" device dbus method call, its on NetworkManager object instead
- Add "setKeyForNetwork" dbus method call on NetworkManager object
* src/NetworkManagerDevice.c
src/NetworkManagerDevice.h
- nm_device_update_link_active(): revert changes for wireless link detection, the WEP-key-is-wrong
logic is in device activation now
- nm_device_activate(): for wireless devices, if we can't associate with access point (perhaps
key is wrong) trigger get-user-key pending action
- Implement get-user-key pending action stuff, tie to dbus messages
- Rename nm_device_ap_list_get_ap() -> nm_device_ap_list_get_ap_by_index()
- Add nm_device_ap_list_get_ap_by_essid()
- Instead of copying "best" access points, ref them instead so that the key we set
sticks around
* src/NetworkManagerPolicy.c
- Deal with wrong WEP key, but right access point (and if so, return link_active = TRUE)
- Don't cancel pending actions on a device if its the same device as last iteration
- Only promote pending_device->active_device if activation was successfull
* src/Makefile.am
- Rename nmclienttest->nmtest
* info-daemon/Makefile.am
info-daemon/NetworkManagerInfo.c
info-daemon/NetworkManagerInfo.h
info-daemon/NetworkManagerInfoDbus.c
info-daemon/NetworkManagerInfoDbus.h
info-daemon/passphrase.glade
info-daemon/NetworkManagerInfo.conf
info-daemon/keyring.png
- Import sources for info-daemon, which pops up dialog for passphrase/key when
NetworkManager asks for it, and also will (soon) provide "allowed" access point
lists to NetworkManager by proxying user's GConf
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@16 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-07-19 06:08:52 +00:00
|
|
|
/*
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
* nm_device_set_user_key_for_network
|
2004-07-19 Dan Williams <dcbw@redhat.com>
* Makefile.am
- Add info-daemon directory
* configure.in
- Check for glade libs and headers
- Add info-daemon directory
* src/NetworkManagerAP.c
- nm_ap_new_from_ap(): Fix bug that resulted in an APs encryption status not getting
copied over to the new AP.
* src/NetworkManagerDbus.c
src/NetworkManagerDbus.h
- Deal with nm_device_ap_list_get_ap()->nm_device_ap_list_get_ap_by_index() change
- Remove nm_dbus_signal_need_key_for_network()
- Add disabled code for asynchronous user wep key callbacks
- Add functions for getting, setting, and cancelling user key operations
- Remove "setKeyForNetwork" device dbus method call, its on NetworkManager object instead
- Add "setKeyForNetwork" dbus method call on NetworkManager object
* src/NetworkManagerDevice.c
src/NetworkManagerDevice.h
- nm_device_update_link_active(): revert changes for wireless link detection, the WEP-key-is-wrong
logic is in device activation now
- nm_device_activate(): for wireless devices, if we can't associate with access point (perhaps
key is wrong) trigger get-user-key pending action
- Implement get-user-key pending action stuff, tie to dbus messages
- Rename nm_device_ap_list_get_ap() -> nm_device_ap_list_get_ap_by_index()
- Add nm_device_ap_list_get_ap_by_essid()
- Instead of copying "best" access points, ref them instead so that the key we set
sticks around
* src/NetworkManagerPolicy.c
- Deal with wrong WEP key, but right access point (and if so, return link_active = TRUE)
- Don't cancel pending actions on a device if its the same device as last iteration
- Only promote pending_device->active_device if activation was successfull
* src/Makefile.am
- Rename nmclienttest->nmtest
* info-daemon/Makefile.am
info-daemon/NetworkManagerInfo.c
info-daemon/NetworkManagerInfo.h
info-daemon/NetworkManagerInfoDbus.c
info-daemon/NetworkManagerInfoDbus.h
info-daemon/passphrase.glade
info-daemon/NetworkManagerInfo.conf
info-daemon/keyring.png
- Import sources for info-daemon, which pops up dialog for passphrase/key when
NetworkManager asks for it, and also will (soon) provide "allowed" access point
lists to NetworkManager by proxying user's GConf
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@16 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-07-19 06:08:52 +00:00
|
|
|
*
|
|
|
|
|
* Called upon receipt of a NetworkManagerInfo reply with a
|
|
|
|
|
* user-supplied key.
|
|
|
|
|
*
|
|
|
|
|
*/
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
void nm_device_set_user_key_for_network (NMDevice *dev, NMAccessPointList *invalid_list,
|
2004-10-08 05:22:17 +00:00
|
|
|
unsigned char *network, unsigned char *key,
|
2004-10-21 17:42:14 +00:00
|
|
|
NMEncKeyType enc_method)
|
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
|
|
|
{
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
NMAccessPoint *best_ap;
|
|
|
|
|
const char *cancel_message = "***canceled***";
|
|
|
|
|
|
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
|
|
|
g_return_if_fail (dev != NULL);
|
|
|
|
|
g_return_if_fail (nm_device_is_wireless (dev));
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
g_return_if_fail (network != NULL);
|
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
|
|
|
g_return_if_fail (key != NULL);
|
|
|
|
|
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
/* If the user canceled, mark the ap as invalid */
|
|
|
|
|
if (strncmp (key, cancel_message, strlen (cancel_message)) == 0)
|
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
|
|
|
{
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
NMAccessPoint *ap;
|
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
|
|
|
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
if ((ap = nm_device_ap_list_get_ap_by_essid (dev, network)))
|
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
|
|
|
{
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
NMAccessPoint *invalid_ap = nm_ap_new_from_ap (ap);
|
|
|
|
|
if (invalid_list)
|
|
|
|
|
nm_ap_list_append_ap (invalid_list, invalid_ap);
|
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
|
|
|
}
|
2004-07-19 Dan Williams <dcbw@redhat.com>
* Makefile.am
- Add info-daemon directory
* configure.in
- Check for glade libs and headers
- Add info-daemon directory
* src/NetworkManagerAP.c
- nm_ap_new_from_ap(): Fix bug that resulted in an APs encryption status not getting
copied over to the new AP.
* src/NetworkManagerDbus.c
src/NetworkManagerDbus.h
- Deal with nm_device_ap_list_get_ap()->nm_device_ap_list_get_ap_by_index() change
- Remove nm_dbus_signal_need_key_for_network()
- Add disabled code for asynchronous user wep key callbacks
- Add functions for getting, setting, and cancelling user key operations
- Remove "setKeyForNetwork" device dbus method call, its on NetworkManager object instead
- Add "setKeyForNetwork" dbus method call on NetworkManager object
* src/NetworkManagerDevice.c
src/NetworkManagerDevice.h
- nm_device_update_link_active(): revert changes for wireless link detection, the WEP-key-is-wrong
logic is in device activation now
- nm_device_activate(): for wireless devices, if we can't associate with access point (perhaps
key is wrong) trigger get-user-key pending action
- Implement get-user-key pending action stuff, tie to dbus messages
- Rename nm_device_ap_list_get_ap() -> nm_device_ap_list_get_ap_by_index()
- Add nm_device_ap_list_get_ap_by_essid()
- Instead of copying "best" access points, ref them instead so that the key we set
sticks around
* src/NetworkManagerPolicy.c
- Deal with wrong WEP key, but right access point (and if so, return link_active = TRUE)
- Don't cancel pending actions on a device if its the same device as last iteration
- Only promote pending_device->active_device if activation was successfull
* src/Makefile.am
- Rename nmclienttest->nmtest
* info-daemon/Makefile.am
info-daemon/NetworkManagerInfo.c
info-daemon/NetworkManagerInfo.h
info-daemon/NetworkManagerInfoDbus.c
info-daemon/NetworkManagerInfoDbus.h
info-daemon/passphrase.glade
info-daemon/NetworkManagerInfo.conf
info-daemon/keyring.png
- Import sources for info-daemon, which pops up dialog for passphrase/key when
NetworkManager asks for it, and also will (soon) provide "allowed" access point
lists to NetworkManager by proxying user's GConf
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@16 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-07-19 06:08:52 +00:00
|
|
|
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
nm_device_update_best_ap (dev);
|
|
|
|
|
}
|
|
|
|
|
else if ((best_ap = nm_device_get_best_ap (dev)))
|
2004-07-19 Dan Williams <dcbw@redhat.com>
* Makefile.am
- Add info-daemon directory
* configure.in
- Check for glade libs and headers
- Add info-daemon directory
* src/NetworkManagerAP.c
- nm_ap_new_from_ap(): Fix bug that resulted in an APs encryption status not getting
copied over to the new AP.
* src/NetworkManagerDbus.c
src/NetworkManagerDbus.h
- Deal with nm_device_ap_list_get_ap()->nm_device_ap_list_get_ap_by_index() change
- Remove nm_dbus_signal_need_key_for_network()
- Add disabled code for asynchronous user wep key callbacks
- Add functions for getting, setting, and cancelling user key operations
- Remove "setKeyForNetwork" device dbus method call, its on NetworkManager object instead
- Add "setKeyForNetwork" dbus method call on NetworkManager object
* src/NetworkManagerDevice.c
src/NetworkManagerDevice.h
- nm_device_update_link_active(): revert changes for wireless link detection, the WEP-key-is-wrong
logic is in device activation now
- nm_device_activate(): for wireless devices, if we can't associate with access point (perhaps
key is wrong) trigger get-user-key pending action
- Implement get-user-key pending action stuff, tie to dbus messages
- Rename nm_device_ap_list_get_ap() -> nm_device_ap_list_get_ap_by_index()
- Add nm_device_ap_list_get_ap_by_essid()
- Instead of copying "best" access points, ref them instead so that the key we set
sticks around
* src/NetworkManagerPolicy.c
- Deal with wrong WEP key, but right access point (and if so, return link_active = TRUE)
- Don't cancel pending actions on a device if its the same device as last iteration
- Only promote pending_device->active_device if activation was successfull
* src/Makefile.am
- Rename nmclienttest->nmtest
* info-daemon/Makefile.am
info-daemon/NetworkManagerInfo.c
info-daemon/NetworkManagerInfo.h
info-daemon/NetworkManagerInfoDbus.c
info-daemon/NetworkManagerInfoDbus.h
info-daemon/passphrase.glade
info-daemon/NetworkManagerInfo.conf
info-daemon/keyring.png
- Import sources for info-daemon, which pops up dialog for passphrase/key when
NetworkManager asks for it, and also will (soon) provide "allowed" access point
lists to NetworkManager by proxying user's GConf
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@16 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-07-19 06:08:52 +00:00
|
|
|
{
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
/* Make sure the "best" ap matches the essid we asked for the key of,
|
|
|
|
|
* then set the new key on the access point.
|
|
|
|
|
*/
|
2004-08-20 18:18:37 +00:00
|
|
|
if (nm_null_safe_strcmp (network, nm_ap_get_essid (best_ap)) == 0)
|
2004-10-08 05:22:17 +00:00
|
|
|
nm_ap_set_enc_key_source (best_ap, key, enc_method);
|
2004-10-29 16:37:43 +00:00
|
|
|
|
|
|
|
|
nm_ap_unref (best_ap);
|
2004-07-19 Dan Williams <dcbw@redhat.com>
* Makefile.am
- Add info-daemon directory
* configure.in
- Check for glade libs and headers
- Add info-daemon directory
* src/NetworkManagerAP.c
- nm_ap_new_from_ap(): Fix bug that resulted in an APs encryption status not getting
copied over to the new AP.
* src/NetworkManagerDbus.c
src/NetworkManagerDbus.h
- Deal with nm_device_ap_list_get_ap()->nm_device_ap_list_get_ap_by_index() change
- Remove nm_dbus_signal_need_key_for_network()
- Add disabled code for asynchronous user wep key callbacks
- Add functions for getting, setting, and cancelling user key operations
- Remove "setKeyForNetwork" device dbus method call, its on NetworkManager object instead
- Add "setKeyForNetwork" dbus method call on NetworkManager object
* src/NetworkManagerDevice.c
src/NetworkManagerDevice.h
- nm_device_update_link_active(): revert changes for wireless link detection, the WEP-key-is-wrong
logic is in device activation now
- nm_device_activate(): for wireless devices, if we can't associate with access point (perhaps
key is wrong) trigger get-user-key pending action
- Implement get-user-key pending action stuff, tie to dbus messages
- Rename nm_device_ap_list_get_ap() -> nm_device_ap_list_get_ap_by_index()
- Add nm_device_ap_list_get_ap_by_essid()
- Instead of copying "best" access points, ref them instead so that the key we set
sticks around
* src/NetworkManagerPolicy.c
- Deal with wrong WEP key, but right access point (and if so, return link_active = TRUE)
- Don't cancel pending actions on a device if its the same device as last iteration
- Only promote pending_device->active_device if activation was successfull
* src/Makefile.am
- Rename nmclienttest->nmtest
* info-daemon/Makefile.am
info-daemon/NetworkManagerInfo.c
info-daemon/NetworkManagerInfo.h
info-daemon/NetworkManagerInfoDbus.c
info-daemon/NetworkManagerInfoDbus.h
info-daemon/passphrase.glade
info-daemon/NetworkManagerInfo.conf
info-daemon/keyring.png
- Import sources for info-daemon, which pops up dialog for passphrase/key when
NetworkManager asks for it, and also will (soon) provide "allowed" access point
lists to NetworkManager by proxying user's GConf
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@16 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-07-19 06:08:52 +00:00
|
|
|
}
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
dev->options.wireless.user_key_received = TRUE;
|
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
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
2004-06-24 14:18:37 +00:00
|
|
|
/*
|
2004-08-31 Dan Williams <dcbw@redhat.com>
* Remove 'debug' extern global from all files since we now
use syslog()
* src/NetworkManager.[ch]
- Break out routine that get the net.interface property from HAL,
removing that logic from nm_create_device_and_add_to_list()
- (nm_create_device_and_add_to_list): make this a bit more general so
it doesn't do the talking to HAL. Also add arguments to facilitate
the create of test devices.
- (nm_data_mark_state_changed): rename from nm_data_set_state_modified()
- (nm_data_new, main, nm_print_usage): add new argument "--enable-test-devices"
which makes NetworkManager listen for dbus commands to create test
devices, which have no backing hardware. Use when you're on a plane
for example, and/or forgot your wireless card at home. Test devices
_cannot_ be created unless NM is started with --enable-test-devices.
* src/NetworkManagerDbus.[ch]
- New "getLinkActive" method for devices
- New "setLinkActive" method for devices (only works on test devices)
- New "createTestDevice" method on NetworkManager object to create a test
device of a specified type (ie wired, wireless). UDI is created from
scratch, as is the interface name. Only works when NM is started with
--enable-test-devices switch.
- New "removeTestDevice" method on NetworkManager object which removes a
test device. Only works when NM is started with --enable-test-devices
* src/NetworkManagerDevice.[ch]
- Logic to facilitate test devices. Add variables to NMDevice struct to indicate
whether a device is a test device or not, and what its link status is.
- Deal with test devices in most functions. For those that work directly on hardware
special-case test devices.
- (nm_device_new): don't create a test device if test devices weren't enabled on the
command-line.
- (nm_device_update_link_active): split out logic for wired and wireless device link
checking to separate functions to facilitate test device link checking.
- (nm_device_set_enc_key): Since some drivers for wireless cards are daft and
don't make a distinction between System Authentication and Encryption
(namely Cisco aironet), we use Open System auth when setting a WEP key
on the card. We don't deal with Shared Key auth yet.
- (nm_device_activation_worker): split the activation cancel check logic out into
a separate routine nm_device_activation_cancel_if_needed()
- (nm_device_activation_signal_cancel): rename from nm_device_activation_cancel()
- (nm_device_fake_ap_list): Test wireless devices obviously cannot scan, so create
a list of fake access points that they can "see"
- (nm_device_is_test_device): return whether or not a device is a test device
* src/NetworkManagerPolicy.c
- (nm_policy_get_best_device): attempt to deal with wireless network selection,
previously if you "locked"/forced NM to use a wireless device but then
selected a wireless network for NM to use, it would switch to a wired device.
So, if the active device is wireless and it has a "forced" best AP, use it
if the "forced" best AP is still valid
- (nm_state_modification_monitor): deal with NULL best devices, for example
there were no usable network devices, or the last one was removed
* src/backends/NetworkManager*.c
- Deal with test devices, mostly just return success for operations like getting
a DHCP address
* test/nmtestdevices.c
- Test tool to create/remove/link-switch test devices
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@112 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-31 16:09:15 +00:00
|
|
|
* nm_device_ap_list_add_ap
|
2004-06-24 14:18:37 +00:00
|
|
|
*
|
|
|
|
|
* Add an access point to the devices internal AP list.
|
|
|
|
|
*
|
|
|
|
|
*/
|
2004-08-31 Dan Williams <dcbw@redhat.com>
* Remove 'debug' extern global from all files since we now
use syslog()
* src/NetworkManager.[ch]
- Break out routine that get the net.interface property from HAL,
removing that logic from nm_create_device_and_add_to_list()
- (nm_create_device_and_add_to_list): make this a bit more general so
it doesn't do the talking to HAL. Also add arguments to facilitate
the create of test devices.
- (nm_data_mark_state_changed): rename from nm_data_set_state_modified()
- (nm_data_new, main, nm_print_usage): add new argument "--enable-test-devices"
which makes NetworkManager listen for dbus commands to create test
devices, which have no backing hardware. Use when you're on a plane
for example, and/or forgot your wireless card at home. Test devices
_cannot_ be created unless NM is started with --enable-test-devices.
* src/NetworkManagerDbus.[ch]
- New "getLinkActive" method for devices
- New "setLinkActive" method for devices (only works on test devices)
- New "createTestDevice" method on NetworkManager object to create a test
device of a specified type (ie wired, wireless). UDI is created from
scratch, as is the interface name. Only works when NM is started with
--enable-test-devices switch.
- New "removeTestDevice" method on NetworkManager object which removes a
test device. Only works when NM is started with --enable-test-devices
* src/NetworkManagerDevice.[ch]
- Logic to facilitate test devices. Add variables to NMDevice struct to indicate
whether a device is a test device or not, and what its link status is.
- Deal with test devices in most functions. For those that work directly on hardware
special-case test devices.
- (nm_device_new): don't create a test device if test devices weren't enabled on the
command-line.
- (nm_device_update_link_active): split out logic for wired and wireless device link
checking to separate functions to facilitate test device link checking.
- (nm_device_set_enc_key): Since some drivers for wireless cards are daft and
don't make a distinction between System Authentication and Encryption
(namely Cisco aironet), we use Open System auth when setting a WEP key
on the card. We don't deal with Shared Key auth yet.
- (nm_device_activation_worker): split the activation cancel check logic out into
a separate routine nm_device_activation_cancel_if_needed()
- (nm_device_activation_signal_cancel): rename from nm_device_activation_cancel()
- (nm_device_fake_ap_list): Test wireless devices obviously cannot scan, so create
a list of fake access points that they can "see"
- (nm_device_is_test_device): return whether or not a device is a test device
* src/NetworkManagerPolicy.c
- (nm_policy_get_best_device): attempt to deal with wireless network selection,
previously if you "locked"/forced NM to use a wireless device but then
selected a wireless network for NM to use, it would switch to a wired device.
So, if the active device is wireless and it has a "forced" best AP, use it
if the "forced" best AP is still valid
- (nm_state_modification_monitor): deal with NULL best devices, for example
there were no usable network devices, or the last one was removed
* src/backends/NetworkManager*.c
- Deal with test devices, mostly just return success for operations like getting
a DHCP address
* test/nmtestdevices.c
- Test tool to create/remove/link-switch test devices
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@112 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-31 16:09:15 +00:00
|
|
|
static void nm_device_ap_list_add_ap (NMDevice *dev, NMAccessPoint *ap)
|
2004-06-24 14:18:37 +00:00
|
|
|
{
|
|
|
|
|
g_return_if_fail (dev != NULL);
|
|
|
|
|
g_return_if_fail (ap != NULL);
|
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
|
|
|
g_return_if_fail (nm_device_is_wireless (dev));
|
2004-06-24 14:18:37 +00:00
|
|
|
|
2004-08-05 18:54:29 +00:00
|
|
|
nm_ap_list_append_ap (dev->options.wireless.ap_list, ap);
|
|
|
|
|
/* Transfer ownership of ap to the list by unrefing it here */
|
|
|
|
|
nm_ap_unref (ap);
|
2004-06-24 14:18:37 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* nm_device_ap_list_clear
|
|
|
|
|
*
|
|
|
|
|
* Clears out the device's internal list of available access points.
|
|
|
|
|
*
|
|
|
|
|
*/
|
|
|
|
|
void nm_device_ap_list_clear (NMDevice *dev)
|
|
|
|
|
{
|
|
|
|
|
g_return_if_fail (dev != NULL);
|
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
|
|
|
g_return_if_fail (nm_device_is_wireless (dev));
|
2004-06-24 14:18:37 +00:00
|
|
|
|
2004-08-05 18:54:29 +00:00
|
|
|
if (!dev->options.wireless.ap_list)
|
2004-06-24 14:18:37 +00:00
|
|
|
return;
|
|
|
|
|
|
2004-08-05 18:54:29 +00:00
|
|
|
nm_ap_list_unref (dev->options.wireless.ap_list);
|
|
|
|
|
dev->options.wireless.ap_list = NULL;
|
2004-06-24 14:18:37 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
2004-07-19 Dan Williams <dcbw@redhat.com>
* Makefile.am
- Add info-daemon directory
* configure.in
- Check for glade libs and headers
- Add info-daemon directory
* src/NetworkManagerAP.c
- nm_ap_new_from_ap(): Fix bug that resulted in an APs encryption status not getting
copied over to the new AP.
* src/NetworkManagerDbus.c
src/NetworkManagerDbus.h
- Deal with nm_device_ap_list_get_ap()->nm_device_ap_list_get_ap_by_index() change
- Remove nm_dbus_signal_need_key_for_network()
- Add disabled code for asynchronous user wep key callbacks
- Add functions for getting, setting, and cancelling user key operations
- Remove "setKeyForNetwork" device dbus method call, its on NetworkManager object instead
- Add "setKeyForNetwork" dbus method call on NetworkManager object
* src/NetworkManagerDevice.c
src/NetworkManagerDevice.h
- nm_device_update_link_active(): revert changes for wireless link detection, the WEP-key-is-wrong
logic is in device activation now
- nm_device_activate(): for wireless devices, if we can't associate with access point (perhaps
key is wrong) trigger get-user-key pending action
- Implement get-user-key pending action stuff, tie to dbus messages
- Rename nm_device_ap_list_get_ap() -> nm_device_ap_list_get_ap_by_index()
- Add nm_device_ap_list_get_ap_by_essid()
- Instead of copying "best" access points, ref them instead so that the key we set
sticks around
* src/NetworkManagerPolicy.c
- Deal with wrong WEP key, but right access point (and if so, return link_active = TRUE)
- Don't cancel pending actions on a device if its the same device as last iteration
- Only promote pending_device->active_device if activation was successfull
* src/Makefile.am
- Rename nmclienttest->nmtest
* info-daemon/Makefile.am
info-daemon/NetworkManagerInfo.c
info-daemon/NetworkManagerInfo.h
info-daemon/NetworkManagerInfoDbus.c
info-daemon/NetworkManagerInfoDbus.h
info-daemon/passphrase.glade
info-daemon/NetworkManagerInfo.conf
info-daemon/keyring.png
- Import sources for info-daemon, which pops up dialog for passphrase/key when
NetworkManager asks for it, and also will (soon) provide "allowed" access point
lists to NetworkManager by proxying user's GConf
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@16 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-07-19 06:08:52 +00:00
|
|
|
/*
|
|
|
|
|
* nm_device_ap_list_get_ap_by_essid
|
|
|
|
|
*
|
|
|
|
|
* Get the access point for a specific essid
|
|
|
|
|
*
|
|
|
|
|
*/
|
|
|
|
|
NMAccessPoint *nm_device_ap_list_get_ap_by_essid (NMDevice *dev, const char *essid)
|
|
|
|
|
{
|
|
|
|
|
NMAccessPoint *ret_ap = NULL;
|
|
|
|
|
|
|
|
|
|
g_return_val_if_fail (dev != NULL, NULL);
|
|
|
|
|
g_return_val_if_fail (nm_device_is_wireless (dev), NULL);
|
|
|
|
|
g_return_val_if_fail (essid != NULL, NULL);
|
|
|
|
|
|
2004-08-05 18:54:29 +00:00
|
|
|
if (!dev->options.wireless.ap_list)
|
2004-07-19 Dan Williams <dcbw@redhat.com>
* Makefile.am
- Add info-daemon directory
* configure.in
- Check for glade libs and headers
- Add info-daemon directory
* src/NetworkManagerAP.c
- nm_ap_new_from_ap(): Fix bug that resulted in an APs encryption status not getting
copied over to the new AP.
* src/NetworkManagerDbus.c
src/NetworkManagerDbus.h
- Deal with nm_device_ap_list_get_ap()->nm_device_ap_list_get_ap_by_index() change
- Remove nm_dbus_signal_need_key_for_network()
- Add disabled code for asynchronous user wep key callbacks
- Add functions for getting, setting, and cancelling user key operations
- Remove "setKeyForNetwork" device dbus method call, its on NetworkManager object instead
- Add "setKeyForNetwork" dbus method call on NetworkManager object
* src/NetworkManagerDevice.c
src/NetworkManagerDevice.h
- nm_device_update_link_active(): revert changes for wireless link detection, the WEP-key-is-wrong
logic is in device activation now
- nm_device_activate(): for wireless devices, if we can't associate with access point (perhaps
key is wrong) trigger get-user-key pending action
- Implement get-user-key pending action stuff, tie to dbus messages
- Rename nm_device_ap_list_get_ap() -> nm_device_ap_list_get_ap_by_index()
- Add nm_device_ap_list_get_ap_by_essid()
- Instead of copying "best" access points, ref them instead so that the key we set
sticks around
* src/NetworkManagerPolicy.c
- Deal with wrong WEP key, but right access point (and if so, return link_active = TRUE)
- Don't cancel pending actions on a device if its the same device as last iteration
- Only promote pending_device->active_device if activation was successfull
* src/Makefile.am
- Rename nmclienttest->nmtest
* info-daemon/Makefile.am
info-daemon/NetworkManagerInfo.c
info-daemon/NetworkManagerInfo.h
info-daemon/NetworkManagerInfoDbus.c
info-daemon/NetworkManagerInfoDbus.h
info-daemon/passphrase.glade
info-daemon/NetworkManagerInfo.conf
info-daemon/keyring.png
- Import sources for info-daemon, which pops up dialog for passphrase/key when
NetworkManager asks for it, and also will (soon) provide "allowed" access point
lists to NetworkManager by proxying user's GConf
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@16 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-07-19 06:08:52 +00:00
|
|
|
return (NULL);
|
|
|
|
|
|
2004-08-05 18:54:29 +00:00
|
|
|
ret_ap = nm_ap_list_get_ap_by_essid (dev->options.wireless.ap_list, essid);
|
2004-07-19 Dan Williams <dcbw@redhat.com>
* Makefile.am
- Add info-daemon directory
* configure.in
- Check for glade libs and headers
- Add info-daemon directory
* src/NetworkManagerAP.c
- nm_ap_new_from_ap(): Fix bug that resulted in an APs encryption status not getting
copied over to the new AP.
* src/NetworkManagerDbus.c
src/NetworkManagerDbus.h
- Deal with nm_device_ap_list_get_ap()->nm_device_ap_list_get_ap_by_index() change
- Remove nm_dbus_signal_need_key_for_network()
- Add disabled code for asynchronous user wep key callbacks
- Add functions for getting, setting, and cancelling user key operations
- Remove "setKeyForNetwork" device dbus method call, its on NetworkManager object instead
- Add "setKeyForNetwork" dbus method call on NetworkManager object
* src/NetworkManagerDevice.c
src/NetworkManagerDevice.h
- nm_device_update_link_active(): revert changes for wireless link detection, the WEP-key-is-wrong
logic is in device activation now
- nm_device_activate(): for wireless devices, if we can't associate with access point (perhaps
key is wrong) trigger get-user-key pending action
- Implement get-user-key pending action stuff, tie to dbus messages
- Rename nm_device_ap_list_get_ap() -> nm_device_ap_list_get_ap_by_index()
- Add nm_device_ap_list_get_ap_by_essid()
- Instead of copying "best" access points, ref them instead so that the key we set
sticks around
* src/NetworkManagerPolicy.c
- Deal with wrong WEP key, but right access point (and if so, return link_active = TRUE)
- Don't cancel pending actions on a device if its the same device as last iteration
- Only promote pending_device->active_device if activation was successfull
* src/Makefile.am
- Rename nmclienttest->nmtest
* info-daemon/Makefile.am
info-daemon/NetworkManagerInfo.c
info-daemon/NetworkManagerInfo.h
info-daemon/NetworkManagerInfoDbus.c
info-daemon/NetworkManagerInfoDbus.h
info-daemon/passphrase.glade
info-daemon/NetworkManagerInfo.conf
info-daemon/keyring.png
- Import sources for info-daemon, which pops up dialog for passphrase/key when
NetworkManager asks for it, and also will (soon) provide "allowed" access point
lists to NetworkManager by proxying user's GConf
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@16 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-07-19 06:08:52 +00:00
|
|
|
|
|
|
|
|
return (ret_ap);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
2004-10-13 20:57:23 +00:00
|
|
|
/*
|
|
|
|
|
* nm_device_ap_list_get_ap_by_address
|
|
|
|
|
*
|
|
|
|
|
* Get the access point for a specific MAC address
|
|
|
|
|
*
|
|
|
|
|
*/
|
|
|
|
|
NMAccessPoint *nm_device_ap_list_get_ap_by_address (NMDevice *dev, const struct ether_addr *addr)
|
|
|
|
|
{
|
|
|
|
|
NMAccessPoint *ret_ap = NULL;
|
|
|
|
|
|
|
|
|
|
g_return_val_if_fail (dev != NULL, NULL);
|
|
|
|
|
g_return_val_if_fail (nm_device_is_wireless (dev), NULL);
|
|
|
|
|
g_return_val_if_fail (addr != NULL, NULL);
|
|
|
|
|
|
|
|
|
|
if (!dev->options.wireless.ap_list)
|
|
|
|
|
return (NULL);
|
|
|
|
|
|
|
|
|
|
ret_ap = nm_ap_list_get_ap_by_address (dev->options.wireless.ap_list, addr);
|
|
|
|
|
|
|
|
|
|
return (ret_ap);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
2004-08-02 21:12:40 +00:00
|
|
|
/*
|
|
|
|
|
* nm_device_ap_list_get
|
|
|
|
|
*
|
|
|
|
|
* Return a pointer to the AP list
|
|
|
|
|
*
|
|
|
|
|
*/
|
2004-08-05 18:54:29 +00:00
|
|
|
NMAccessPointList *nm_device_ap_list_get (NMDevice *dev)
|
2004-08-02 21:12:40 +00:00
|
|
|
{
|
|
|
|
|
g_return_val_if_fail (dev != NULL, NULL);
|
|
|
|
|
g_return_val_if_fail (nm_device_is_wireless (dev), NULL);
|
|
|
|
|
|
2004-08-05 18:54:29 +00:00
|
|
|
return (dev->options.wireless.ap_list);
|
2004-08-02 21:12:40 +00:00
|
|
|
}
|
|
|
|
|
|
2004-07-06 01:34:10 +00:00
|
|
|
/*
|
|
|
|
|
* Get/Set functions for "best" access point
|
|
|
|
|
*
|
2004-10-29 16:37:43 +00:00
|
|
|
* Caller MUST unref returned access point when done with it.
|
|
|
|
|
*
|
2004-07-06 01:34:10 +00:00
|
|
|
*/
|
|
|
|
|
NMAccessPoint *nm_device_get_best_ap (NMDevice *dev)
|
|
|
|
|
{
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
NMAccessPoint *best_ap;
|
|
|
|
|
|
2004-07-06 01:34:10 +00:00
|
|
|
g_return_val_if_fail (dev != NULL, NULL);
|
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
|
|
|
g_return_val_if_fail (nm_device_is_wireless (dev), NULL);
|
2004-07-06 01:34:10 +00:00
|
|
|
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
g_mutex_lock (dev->options.wireless.best_ap_mutex);
|
|
|
|
|
best_ap = dev->options.wireless.best_ap;
|
2004-10-29 16:37:43 +00:00
|
|
|
/* Callers get a reffed AP */
|
|
|
|
|
if (best_ap) nm_ap_ref (best_ap);
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
g_mutex_unlock (dev->options.wireless.best_ap_mutex);
|
|
|
|
|
|
|
|
|
|
return (best_ap);
|
2004-07-06 01:34:10 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
void nm_device_set_best_ap (NMDevice *dev, NMAccessPoint *ap)
|
|
|
|
|
{
|
|
|
|
|
g_return_if_fail (dev != NULL);
|
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
|
|
|
g_return_if_fail (nm_device_is_wireless (dev));
|
2004-07-06 01:34:10 +00:00
|
|
|
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
g_mutex_lock (dev->options.wireless.best_ap_mutex);
|
|
|
|
|
|
2004-08-05 18:54:29 +00:00
|
|
|
if (dev->options.wireless.best_ap)
|
|
|
|
|
nm_ap_unref (dev->options.wireless.best_ap);
|
|
|
|
|
|
|
|
|
|
if (ap)
|
|
|
|
|
nm_ap_ref (ap);
|
|
|
|
|
|
|
|
|
|
dev->options.wireless.best_ap = ap;
|
2004-09-09 20:02:59 +00:00
|
|
|
nm_device_unfreeze_best_ap (dev);
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
g_mutex_unlock (dev->options.wireless.best_ap_mutex);
|
2004-08-05 18:54:29 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
2004-08-12 Dan Williams <dcbw@redhat.com>
* info-daemon/passphrase.glade
- Set window title to " "
* panel-applet/Makefile.am
panel-applet/keyring.png
- Deliver to correct place
* panel-applet/NMWirelessApplet.[ch]
- Add comments
- Remove applet->have_active_device as its no longer used
- (nmwa_load_theme): load keyring.png too
- (error_dialog): remove
- (show_warning_dialog): subsume functionality of error dialog too
- (nmwa_destroy, nmwa_new): create and dispose of an application-wide GConfClient
- (nmwa_handle_network_choice): add to deal with user clicking on an item from
the networks menu
- (nmwa_menu_item_activated): GtkMenuItem "activate" signal handler
- (nmwa_button_clicked, nmwa_setup_widgets): create and populate the menu on startup
and when we get broadcasts of changed wireless access points only, not when the
user clicks on the button to display the menu (too long of a wait)
- (nmwa_add_menu_item): Make active network bold, and place a keyring icon beside
networks that are encrypted
- (nmwa_dispose_menu, nmwa_menu_item_data_free): dispose of the data we place on each
menu item with g_object_set_data()
* panel-applet/NMWirelessAppletDbus.[ch]
- (nmwa_dbus_get_bool): add method to return boolean value from dbus message
- (nmwa_dbus_get_active_network): add (nmwa_dbus_get_string() wrapper to get active network)
- (nmwa_dbus_add_networks_to_menu): clean up, only show one instance of each ESSID in the menu
- (nmwa_dbus_set_network): force NetworkManager to use a particular network for wireless cards
- (nmwa_dbus_init, nmwa_dbus_filter): Trap network appear/disappear and device
activation/deactivation signals and rebuild the menu when they happen
* src/NetworkManager.c
- (main): use new nm_spawn_process() rather than system()
* src/NetworkManagerDbus.c
- (nm_dbus_devices_handle_request): don't compare AP structure addresses directly, but essids
instead. Since we can now force best_aps to stick around, the AP structure to which
dev->options.wireless.best_ap points to won't necessarily be in the device's device list
if a scan has happened since the best_ap was frozen. Also add "setNetwork" method
to freeze the best_ap.
* src/NetworkManagerDevice.[ch]
- (nm_device_activation_worker): Use new nm_spawn_process() call rather than system()
- (nm_device_*_best_ap): add freeze/unfreeze/get_frozen functions, and don't really update
the best_ap in nm_device_update_best_ap() if the best_ap is frozen AND in the device's
ap list
* src/NetworkManagerUtils.[ch]
- (nm_spawn_process): add replacement for system() usage
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@48 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-12 19:58:01 +00:00
|
|
|
/*
|
|
|
|
|
* Freeze/unfreeze best ap
|
|
|
|
|
*
|
|
|
|
|
* If the user explicitly picks a network to associate with, we don't
|
|
|
|
|
* change the active network until it goes out of range.
|
|
|
|
|
*
|
|
|
|
|
*/
|
|
|
|
|
void nm_device_freeze_best_ap (NMDevice *dev)
|
|
|
|
|
{
|
|
|
|
|
g_return_if_fail (dev != NULL);
|
|
|
|
|
g_return_if_fail (nm_device_is_wireless (dev));
|
|
|
|
|
|
|
|
|
|
dev->options.wireless.freeze_best_ap = TRUE;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
void nm_device_unfreeze_best_ap (NMDevice *dev)
|
|
|
|
|
{
|
|
|
|
|
g_return_if_fail (dev != NULL);
|
|
|
|
|
g_return_if_fail (nm_device_is_wireless (dev));
|
|
|
|
|
|
|
|
|
|
dev->options.wireless.freeze_best_ap = FALSE;
|
|
|
|
|
}
|
|
|
|
|
|
2004-10-11 21:32:19 +00:00
|
|
|
gboolean nm_device_is_best_ap_frozen (NMDevice *dev)
|
2004-08-12 Dan Williams <dcbw@redhat.com>
* info-daemon/passphrase.glade
- Set window title to " "
* panel-applet/Makefile.am
panel-applet/keyring.png
- Deliver to correct place
* panel-applet/NMWirelessApplet.[ch]
- Add comments
- Remove applet->have_active_device as its no longer used
- (nmwa_load_theme): load keyring.png too
- (error_dialog): remove
- (show_warning_dialog): subsume functionality of error dialog too
- (nmwa_destroy, nmwa_new): create and dispose of an application-wide GConfClient
- (nmwa_handle_network_choice): add to deal with user clicking on an item from
the networks menu
- (nmwa_menu_item_activated): GtkMenuItem "activate" signal handler
- (nmwa_button_clicked, nmwa_setup_widgets): create and populate the menu on startup
and when we get broadcasts of changed wireless access points only, not when the
user clicks on the button to display the menu (too long of a wait)
- (nmwa_add_menu_item): Make active network bold, and place a keyring icon beside
networks that are encrypted
- (nmwa_dispose_menu, nmwa_menu_item_data_free): dispose of the data we place on each
menu item with g_object_set_data()
* panel-applet/NMWirelessAppletDbus.[ch]
- (nmwa_dbus_get_bool): add method to return boolean value from dbus message
- (nmwa_dbus_get_active_network): add (nmwa_dbus_get_string() wrapper to get active network)
- (nmwa_dbus_add_networks_to_menu): clean up, only show one instance of each ESSID in the menu
- (nmwa_dbus_set_network): force NetworkManager to use a particular network for wireless cards
- (nmwa_dbus_init, nmwa_dbus_filter): Trap network appear/disappear and device
activation/deactivation signals and rebuild the menu when they happen
* src/NetworkManager.c
- (main): use new nm_spawn_process() rather than system()
* src/NetworkManagerDbus.c
- (nm_dbus_devices_handle_request): don't compare AP structure addresses directly, but essids
instead. Since we can now force best_aps to stick around, the AP structure to which
dev->options.wireless.best_ap points to won't necessarily be in the device's device list
if a scan has happened since the best_ap was frozen. Also add "setNetwork" method
to freeze the best_ap.
* src/NetworkManagerDevice.[ch]
- (nm_device_activation_worker): Use new nm_spawn_process() call rather than system()
- (nm_device_*_best_ap): add freeze/unfreeze/get_frozen functions, and don't really update
the best_ap in nm_device_update_best_ap() if the best_ap is frozen AND in the device's
ap list
* src/NetworkManagerUtils.[ch]
- (nm_spawn_process): add replacement for system() usage
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@48 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-12 19:58:01 +00:00
|
|
|
{
|
|
|
|
|
g_return_val_if_fail (dev != NULL, FALSE);
|
|
|
|
|
g_return_val_if_fail (nm_device_is_wireless (dev), FALSE);
|
|
|
|
|
|
|
|
|
|
return (dev->options.wireless.freeze_best_ap);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
/*
|
|
|
|
|
* nm_device_get_path_for_ap
|
|
|
|
|
*
|
|
|
|
|
* Return the object path for an access point.
|
|
|
|
|
*
|
|
|
|
|
* NOTE: assumes the access point is actually in the device's access point list.
|
|
|
|
|
*
|
|
|
|
|
*/
|
2004-08-05 18:54:29 +00:00
|
|
|
char * nm_device_get_path_for_ap (NMDevice *dev, NMAccessPoint *ap)
|
|
|
|
|
{
|
|
|
|
|
g_return_val_if_fail (dev != NULL, NULL);
|
|
|
|
|
g_return_val_if_fail (ap != NULL, NULL);
|
|
|
|
|
|
2004-10-13 20:57:23 +00:00
|
|
|
if (nm_ap_get_essid (ap))
|
|
|
|
|
return (g_strdup_printf ("%s/%s/Networks/%s", NM_DBUS_PATH_DEVICES, nm_device_get_iface (dev), nm_ap_get_essid (ap)));
|
|
|
|
|
else
|
|
|
|
|
return (NULL);
|
2004-07-06 01:34:10 +00:00
|
|
|
}
|
|
|
|
|
|
2004-08-05 18:54:29 +00:00
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* nm_device_need_ap_switch
|
|
|
|
|
*
|
|
|
|
|
* Returns TRUE if the essid of the card does not match the essid
|
|
|
|
|
* of the "best" access point it should be associating with.
|
|
|
|
|
*
|
|
|
|
|
*/
|
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
|
|
|
gboolean nm_device_need_ap_switch (NMDevice *dev)
|
|
|
|
|
{
|
|
|
|
|
NMAccessPoint *ap;
|
|
|
|
|
gboolean need_switch = FALSE;
|
|
|
|
|
|
|
|
|
|
g_return_val_if_fail (dev != NULL, FALSE);
|
|
|
|
|
g_return_val_if_fail (nm_device_is_wireless (dev), FALSE);
|
|
|
|
|
|
|
|
|
|
ap = nm_device_get_best_ap (dev);
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
if (nm_null_safe_strcmp (nm_device_get_essid (dev), (ap ? nm_ap_get_essid (ap) : NULL)) != 0)
|
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
|
|
|
need_switch = TRUE;
|
|
|
|
|
|
2004-10-29 16:37:43 +00:00
|
|
|
if (ap) nm_ap_unref (ap);
|
|
|
|
|
|
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
|
|
|
return (need_switch);
|
|
|
|
|
}
|
2004-07-06 01:34:10 +00:00
|
|
|
|
2004-08-05 18:54:29 +00:00
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* nm_device_update_best_ap
|
|
|
|
|
*
|
2004-09-09 20:02:59 +00:00
|
|
|
* Recalculate the "best" access point we should be associating with. This
|
|
|
|
|
* function may disrupt the current connection, so it should be called only
|
|
|
|
|
* when necessary, ie when the current access point is no longer in range
|
|
|
|
|
* or is for some other reason invalid and should no longer be used.
|
2004-08-05 18:54:29 +00:00
|
|
|
*
|
|
|
|
|
*/
|
|
|
|
|
void nm_device_update_best_ap (NMDevice *dev)
|
|
|
|
|
{
|
|
|
|
|
NMAccessPointList *ap_list;
|
|
|
|
|
NMAPListIter *iter;
|
2004-10-29 16:37:43 +00:00
|
|
|
NMAccessPoint *scan_ap = NULL;
|
2004-08-05 18:54:29 +00:00
|
|
|
NMAccessPoint *best_ap = NULL;
|
2004-09-08 18:14:42 +00:00
|
|
|
NMAccessPoint *trusted_best_ap = NULL;
|
|
|
|
|
NMAccessPoint *untrusted_best_ap = NULL;
|
|
|
|
|
GTimeVal trusted_latest_timestamp = {0, 0};
|
|
|
|
|
GTimeVal untrusted_latest_timestamp = {0, 0};
|
2004-08-05 18:54:29 +00:00
|
|
|
|
|
|
|
|
g_return_if_fail (dev != NULL);
|
|
|
|
|
g_return_if_fail (dev->app_data != NULL);
|
|
|
|
|
g_return_if_fail (nm_device_is_wireless (dev));
|
|
|
|
|
|
|
|
|
|
if (!(ap_list = nm_device_ap_list_get (dev)))
|
|
|
|
|
return;
|
|
|
|
|
|
2004-08-12 Dan Williams <dcbw@redhat.com>
* info-daemon/passphrase.glade
- Set window title to " "
* panel-applet/Makefile.am
panel-applet/keyring.png
- Deliver to correct place
* panel-applet/NMWirelessApplet.[ch]
- Add comments
- Remove applet->have_active_device as its no longer used
- (nmwa_load_theme): load keyring.png too
- (error_dialog): remove
- (show_warning_dialog): subsume functionality of error dialog too
- (nmwa_destroy, nmwa_new): create and dispose of an application-wide GConfClient
- (nmwa_handle_network_choice): add to deal with user clicking on an item from
the networks menu
- (nmwa_menu_item_activated): GtkMenuItem "activate" signal handler
- (nmwa_button_clicked, nmwa_setup_widgets): create and populate the menu on startup
and when we get broadcasts of changed wireless access points only, not when the
user clicks on the button to display the menu (too long of a wait)
- (nmwa_add_menu_item): Make active network bold, and place a keyring icon beside
networks that are encrypted
- (nmwa_dispose_menu, nmwa_menu_item_data_free): dispose of the data we place on each
menu item with g_object_set_data()
* panel-applet/NMWirelessAppletDbus.[ch]
- (nmwa_dbus_get_bool): add method to return boolean value from dbus message
- (nmwa_dbus_get_active_network): add (nmwa_dbus_get_string() wrapper to get active network)
- (nmwa_dbus_add_networks_to_menu): clean up, only show one instance of each ESSID in the menu
- (nmwa_dbus_set_network): force NetworkManager to use a particular network for wireless cards
- (nmwa_dbus_init, nmwa_dbus_filter): Trap network appear/disappear and device
activation/deactivation signals and rebuild the menu when they happen
* src/NetworkManager.c
- (main): use new nm_spawn_process() rather than system()
* src/NetworkManagerDbus.c
- (nm_dbus_devices_handle_request): don't compare AP structure addresses directly, but essids
instead. Since we can now force best_aps to stick around, the AP structure to which
dev->options.wireless.best_ap points to won't necessarily be in the device's device list
if a scan has happened since the best_ap was frozen. Also add "setNetwork" method
to freeze the best_ap.
* src/NetworkManagerDevice.[ch]
- (nm_device_activation_worker): Use new nm_spawn_process() call rather than system()
- (nm_device_*_best_ap): add freeze/unfreeze/get_frozen functions, and don't really update
the best_ap in nm_device_update_best_ap() if the best_ap is frozen AND in the device's
ap list
* src/NetworkManagerUtils.[ch]
- (nm_spawn_process): add replacement for system() usage
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@48 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-12 19:58:01 +00:00
|
|
|
/* Iterate over the device's ap list to make sure the current
|
|
|
|
|
* "best" ap is still in the device's ap list (so that if its
|
|
|
|
|
* not, we can "unfreeze" the best ap if its been frozen already).
|
|
|
|
|
* If it is, we don't change the best ap here.
|
|
|
|
|
*/
|
2004-10-11 21:32:19 +00:00
|
|
|
if (nm_device_is_best_ap_frozen (dev))
|
2004-08-12 Dan Williams <dcbw@redhat.com>
* info-daemon/passphrase.glade
- Set window title to " "
* panel-applet/Makefile.am
panel-applet/keyring.png
- Deliver to correct place
* panel-applet/NMWirelessApplet.[ch]
- Add comments
- Remove applet->have_active_device as its no longer used
- (nmwa_load_theme): load keyring.png too
- (error_dialog): remove
- (show_warning_dialog): subsume functionality of error dialog too
- (nmwa_destroy, nmwa_new): create and dispose of an application-wide GConfClient
- (nmwa_handle_network_choice): add to deal with user clicking on an item from
the networks menu
- (nmwa_menu_item_activated): GtkMenuItem "activate" signal handler
- (nmwa_button_clicked, nmwa_setup_widgets): create and populate the menu on startup
and when we get broadcasts of changed wireless access points only, not when the
user clicks on the button to display the menu (too long of a wait)
- (nmwa_add_menu_item): Make active network bold, and place a keyring icon beside
networks that are encrypted
- (nmwa_dispose_menu, nmwa_menu_item_data_free): dispose of the data we place on each
menu item with g_object_set_data()
* panel-applet/NMWirelessAppletDbus.[ch]
- (nmwa_dbus_get_bool): add method to return boolean value from dbus message
- (nmwa_dbus_get_active_network): add (nmwa_dbus_get_string() wrapper to get active network)
- (nmwa_dbus_add_networks_to_menu): clean up, only show one instance of each ESSID in the menu
- (nmwa_dbus_set_network): force NetworkManager to use a particular network for wireless cards
- (nmwa_dbus_init, nmwa_dbus_filter): Trap network appear/disappear and device
activation/deactivation signals and rebuild the menu when they happen
* src/NetworkManager.c
- (main): use new nm_spawn_process() rather than system()
* src/NetworkManagerDbus.c
- (nm_dbus_devices_handle_request): don't compare AP structure addresses directly, but essids
instead. Since we can now force best_aps to stick around, the AP structure to which
dev->options.wireless.best_ap points to won't necessarily be in the device's device list
if a scan has happened since the best_ap was frozen. Also add "setNetwork" method
to freeze the best_ap.
* src/NetworkManagerDevice.[ch]
- (nm_device_activation_worker): Use new nm_spawn_process() call rather than system()
- (nm_device_*_best_ap): add freeze/unfreeze/get_frozen functions, and don't really update
the best_ap in nm_device_update_best_ap() if the best_ap is frozen AND in the device's
ap list
* src/NetworkManagerUtils.[ch]
- (nm_spawn_process): add replacement for system() usage
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@48 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-12 19:58:01 +00:00
|
|
|
{
|
2004-10-29 16:37:43 +00:00
|
|
|
best_ap = nm_device_get_best_ap (dev);
|
2004-08-12 Dan Williams <dcbw@redhat.com>
* info-daemon/passphrase.glade
- Set window title to " "
* panel-applet/Makefile.am
panel-applet/keyring.png
- Deliver to correct place
* panel-applet/NMWirelessApplet.[ch]
- Add comments
- Remove applet->have_active_device as its no longer used
- (nmwa_load_theme): load keyring.png too
- (error_dialog): remove
- (show_warning_dialog): subsume functionality of error dialog too
- (nmwa_destroy, nmwa_new): create and dispose of an application-wide GConfClient
- (nmwa_handle_network_choice): add to deal with user clicking on an item from
the networks menu
- (nmwa_menu_item_activated): GtkMenuItem "activate" signal handler
- (nmwa_button_clicked, nmwa_setup_widgets): create and populate the menu on startup
and when we get broadcasts of changed wireless access points only, not when the
user clicks on the button to display the menu (too long of a wait)
- (nmwa_add_menu_item): Make active network bold, and place a keyring icon beside
networks that are encrypted
- (nmwa_dispose_menu, nmwa_menu_item_data_free): dispose of the data we place on each
menu item with g_object_set_data()
* panel-applet/NMWirelessAppletDbus.[ch]
- (nmwa_dbus_get_bool): add method to return boolean value from dbus message
- (nmwa_dbus_get_active_network): add (nmwa_dbus_get_string() wrapper to get active network)
- (nmwa_dbus_add_networks_to_menu): clean up, only show one instance of each ESSID in the menu
- (nmwa_dbus_set_network): force NetworkManager to use a particular network for wireless cards
- (nmwa_dbus_init, nmwa_dbus_filter): Trap network appear/disappear and device
activation/deactivation signals and rebuild the menu when they happen
* src/NetworkManager.c
- (main): use new nm_spawn_process() rather than system()
* src/NetworkManagerDbus.c
- (nm_dbus_devices_handle_request): don't compare AP structure addresses directly, but essids
instead. Since we can now force best_aps to stick around, the AP structure to which
dev->options.wireless.best_ap points to won't necessarily be in the device's device list
if a scan has happened since the best_ap was frozen. Also add "setNetwork" method
to freeze the best_ap.
* src/NetworkManagerDevice.[ch]
- (nm_device_activation_worker): Use new nm_spawn_process() call rather than system()
- (nm_device_*_best_ap): add freeze/unfreeze/get_frozen functions, and don't really update
the best_ap in nm_device_update_best_ap() if the best_ap is frozen AND in the device's
ap list
* src/NetworkManagerUtils.[ch]
- (nm_spawn_process): add replacement for system() usage
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@48 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-12 19:58:01 +00:00
|
|
|
|
|
|
|
|
/* If its in the device's ap list still, don't change the
|
|
|
|
|
* best ap, since its frozen.
|
|
|
|
|
*/
|
2004-08-24 Dan Williams <dcbw@redhat.com>
* src/NetworkManagerAP.[ch]
- Add a "enc_method_good" member and accessors to an Access Point
to signal when we've found the correct encryption method
for an access point
- Add a "timestamp" member and accessors, remove "priority" member
and accessors (use timestamps instead)
- Rename "wep_key"->"enc_key"
- (nm_ap_get_enc_key_hashed): new, return the correct mangled key
for a specified encryption method using the access points
source encryption key/passphrase
* src/NetworkManagerAPList.c
- When updating a network with dbus, grab timestamp now instead of
priority
* src/NetworkManagerDBus.[ch]
- Add signal for "DeviceActivating"
- Switch priority->timestamp
* src/NetworkManagerDevice.c
- Change references of "wep_key" -> "enc_key" or "key"
- Signal DeviceActivating when starting activation
- When activating a wireless device, if the access point we are connecting
to is encrypted, and we have a source key, try to generate a mangled
key and use that (ie, generate real WEP key from a passphrase)
- Rework device activation to fallback to other encryption methods if
a previous one didn't work (ie, try mangling a key as a 104-bit passphrase
first, then if that doesn't work fall back to direct hex key).
- (nm_device_update_best_ap): fix a deadlock, and use timestamps instead of
priority. We now prefer the latest access point used, rather than using
a priority scheme
- (nm_device_do_normal_scan): make the encryption method "unknown" on access
points we've just discovered, and merge in correct info from the global
access point lists
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@68 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-25 22:41:12 +00:00
|
|
|
g_mutex_lock (dev->options.wireless.best_ap_mutex);
|
2004-10-29 16:37:43 +00:00
|
|
|
if (best_ap)
|
2004-08-12 Dan Williams <dcbw@redhat.com>
* info-daemon/passphrase.glade
- Set window title to " "
* panel-applet/Makefile.am
panel-applet/keyring.png
- Deliver to correct place
* panel-applet/NMWirelessApplet.[ch]
- Add comments
- Remove applet->have_active_device as its no longer used
- (nmwa_load_theme): load keyring.png too
- (error_dialog): remove
- (show_warning_dialog): subsume functionality of error dialog too
- (nmwa_destroy, nmwa_new): create and dispose of an application-wide GConfClient
- (nmwa_handle_network_choice): add to deal with user clicking on an item from
the networks menu
- (nmwa_menu_item_activated): GtkMenuItem "activate" signal handler
- (nmwa_button_clicked, nmwa_setup_widgets): create and populate the menu on startup
and when we get broadcasts of changed wireless access points only, not when the
user clicks on the button to display the menu (too long of a wait)
- (nmwa_add_menu_item): Make active network bold, and place a keyring icon beside
networks that are encrypted
- (nmwa_dispose_menu, nmwa_menu_item_data_free): dispose of the data we place on each
menu item with g_object_set_data()
* panel-applet/NMWirelessAppletDbus.[ch]
- (nmwa_dbus_get_bool): add method to return boolean value from dbus message
- (nmwa_dbus_get_active_network): add (nmwa_dbus_get_string() wrapper to get active network)
- (nmwa_dbus_add_networks_to_menu): clean up, only show one instance of each ESSID in the menu
- (nmwa_dbus_set_network): force NetworkManager to use a particular network for wireless cards
- (nmwa_dbus_init, nmwa_dbus_filter): Trap network appear/disappear and device
activation/deactivation signals and rebuild the menu when they happen
* src/NetworkManager.c
- (main): use new nm_spawn_process() rather than system()
* src/NetworkManagerDbus.c
- (nm_dbus_devices_handle_request): don't compare AP structure addresses directly, but essids
instead. Since we can now force best_aps to stick around, the AP structure to which
dev->options.wireless.best_ap points to won't necessarily be in the device's device list
if a scan has happened since the best_ap was frozen. Also add "setNetwork" method
to freeze the best_ap.
* src/NetworkManagerDevice.[ch]
- (nm_device_activation_worker): Use new nm_spawn_process() call rather than system()
- (nm_device_*_best_ap): add freeze/unfreeze/get_frozen functions, and don't really update
the best_ap in nm_device_update_best_ap() if the best_ap is frozen AND in the device's
ap list
* src/NetworkManagerUtils.[ch]
- (nm_spawn_process): add replacement for system() usage
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@48 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-12 19:58:01 +00:00
|
|
|
{
|
2004-10-29 16:37:43 +00:00
|
|
|
char *essid = nm_ap_get_essid (best_ap);
|
|
|
|
|
if ( !nm_ap_list_get_ap_by_essid (dev->app_data->invalid_ap_list, essid)
|
|
|
|
|
&& nm_device_ap_list_get_ap_by_essid (dev, essid))
|
|
|
|
|
{
|
|
|
|
|
nm_ap_unref (best_ap);
|
|
|
|
|
g_mutex_unlock (dev->options.wireless.best_ap_mutex);
|
|
|
|
|
return;
|
|
|
|
|
}
|
|
|
|
|
nm_ap_unref (best_ap);
|
2004-08-12 Dan Williams <dcbw@redhat.com>
* info-daemon/passphrase.glade
- Set window title to " "
* panel-applet/Makefile.am
panel-applet/keyring.png
- Deliver to correct place
* panel-applet/NMWirelessApplet.[ch]
- Add comments
- Remove applet->have_active_device as its no longer used
- (nmwa_load_theme): load keyring.png too
- (error_dialog): remove
- (show_warning_dialog): subsume functionality of error dialog too
- (nmwa_destroy, nmwa_new): create and dispose of an application-wide GConfClient
- (nmwa_handle_network_choice): add to deal with user clicking on an item from
the networks menu
- (nmwa_menu_item_activated): GtkMenuItem "activate" signal handler
- (nmwa_button_clicked, nmwa_setup_widgets): create and populate the menu on startup
and when we get broadcasts of changed wireless access points only, not when the
user clicks on the button to display the menu (too long of a wait)
- (nmwa_add_menu_item): Make active network bold, and place a keyring icon beside
networks that are encrypted
- (nmwa_dispose_menu, nmwa_menu_item_data_free): dispose of the data we place on each
menu item with g_object_set_data()
* panel-applet/NMWirelessAppletDbus.[ch]
- (nmwa_dbus_get_bool): add method to return boolean value from dbus message
- (nmwa_dbus_get_active_network): add (nmwa_dbus_get_string() wrapper to get active network)
- (nmwa_dbus_add_networks_to_menu): clean up, only show one instance of each ESSID in the menu
- (nmwa_dbus_set_network): force NetworkManager to use a particular network for wireless cards
- (nmwa_dbus_init, nmwa_dbus_filter): Trap network appear/disappear and device
activation/deactivation signals and rebuild the menu when they happen
* src/NetworkManager.c
- (main): use new nm_spawn_process() rather than system()
* src/NetworkManagerDbus.c
- (nm_dbus_devices_handle_request): don't compare AP structure addresses directly, but essids
instead. Since we can now force best_aps to stick around, the AP structure to which
dev->options.wireless.best_ap points to won't necessarily be in the device's device list
if a scan has happened since the best_ap was frozen. Also add "setNetwork" method
to freeze the best_ap.
* src/NetworkManagerDevice.[ch]
- (nm_device_activation_worker): Use new nm_spawn_process() call rather than system()
- (nm_device_*_best_ap): add freeze/unfreeze/get_frozen functions, and don't really update
the best_ap in nm_device_update_best_ap() if the best_ap is frozen AND in the device's
ap list
* src/NetworkManagerUtils.[ch]
- (nm_spawn_process): add replacement for system() usage
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@48 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-12 19:58:01 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* Otherwise, its gone away and we don't care about it anymore */
|
|
|
|
|
nm_device_unfreeze_best_ap (dev);
|
|
|
|
|
g_mutex_unlock (dev->options.wireless.best_ap_mutex);
|
|
|
|
|
}
|
|
|
|
|
|
2004-08-05 18:54:29 +00:00
|
|
|
if (!(iter = nm_ap_list_iter_new (ap_list)))
|
|
|
|
|
return;
|
2004-10-29 16:37:43 +00:00
|
|
|
while ((scan_ap = nm_ap_list_iter_next (iter)))
|
2004-08-05 18:54:29 +00:00
|
|
|
{
|
2004-09-08 18:14:42 +00:00
|
|
|
NMAccessPoint *tmp_ap;
|
2004-10-29 16:37:43 +00:00
|
|
|
char *ap_essid = nm_ap_get_essid (scan_ap);
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
|
2004-08-06 18:19:06 +00:00
|
|
|
/* Access points in the "invalid" list cannot be used */
|
2004-09-08 18:14:42 +00:00
|
|
|
if (nm_ap_list_get_ap_by_essid (dev->app_data->invalid_ap_list, ap_essid))
|
|
|
|
|
continue;
|
|
|
|
|
|
|
|
|
|
if ((tmp_ap = nm_ap_list_get_ap_by_essid (dev->app_data->allowed_ap_list, ap_essid)))
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
{
|
2004-08-29 05:46:59 +00:00
|
|
|
const GTimeVal *curtime = nm_ap_get_timestamp (tmp_ap);
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
|
2004-09-08 18:14:42 +00:00
|
|
|
if (nm_ap_get_trusted (tmp_ap) && (curtime->tv_sec > trusted_latest_timestamp.tv_sec))
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
{
|
2004-09-08 18:14:42 +00:00
|
|
|
trusted_latest_timestamp = *nm_ap_get_timestamp (tmp_ap);
|
2004-10-29 16:37:43 +00:00
|
|
|
trusted_best_ap = scan_ap;
|
2004-09-08 18:14:42 +00:00
|
|
|
/* Merge access point data (mainly to get updated WEP key) */
|
2004-10-08 05:22:17 +00:00
|
|
|
nm_ap_set_enc_key_source (trusted_best_ap, nm_ap_get_enc_key_source (tmp_ap), nm_ap_get_enc_method (tmp_ap));
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
}
|
2004-09-08 18:14:42 +00:00
|
|
|
else if (!nm_ap_get_trusted (tmp_ap) && (curtime->tv_sec > untrusted_latest_timestamp.tv_sec))
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
{
|
2004-09-08 18:14:42 +00:00
|
|
|
untrusted_latest_timestamp = *nm_ap_get_timestamp (tmp_ap);
|
2004-10-29 16:37:43 +00:00
|
|
|
untrusted_best_ap = scan_ap;
|
2004-09-08 18:14:42 +00:00
|
|
|
/* Merge access point data (mainly to get updated WEP key) */
|
2004-10-08 05:22:17 +00:00
|
|
|
nm_ap_set_enc_key_source (untrusted_best_ap, nm_ap_get_enc_key_source (tmp_ap), nm_ap_get_enc_method (tmp_ap));
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
}
|
2004-08-06 18:19:06 +00:00
|
|
|
}
|
|
|
|
|
}
|
2004-09-08 18:14:42 +00:00
|
|
|
best_ap = trusted_best_ap ? trusted_best_ap : untrusted_best_ap;
|
|
|
|
|
nm_ap_list_iter_free (iter);
|
2004-08-06 18:19:06 +00:00
|
|
|
|
2004-08-11 19:31:41 +00:00
|
|
|
/* If the best ap is NULL, bring device down and clear out its essid and AP */
|
2004-08-05 18:54:29 +00:00
|
|
|
nm_device_set_best_ap (dev, best_ap);
|
2004-10-29 16:37:43 +00:00
|
|
|
if (!best_ap)
|
2004-08-11 19:31:41 +00:00
|
|
|
{
|
2004-10-29 19:52:35 +00:00
|
|
|
nm_device_set_essid (dev, " ");
|
2004-11-03 07:47:32 +00:00
|
|
|
nm_device_set_enc_key (dev, NULL, NM_DEVICE_AUTH_METHOD_NONE);
|
2004-08-11 19:31:41 +00:00
|
|
|
nm_device_bring_up (dev);
|
|
|
|
|
}
|
2004-08-05 18:54:29 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
2004-10-13 20:57:23 +00:00
|
|
|
/*
|
|
|
|
|
* nm_device_wireless_network_exists
|
|
|
|
|
*
|
|
|
|
|
* Tell the card to explicitly use with a particular essid, and then
|
|
|
|
|
* see if we can associate with some AP using that ESSID.
|
|
|
|
|
* Mainly for non-essid-broadcasting APs to figure out whether or not
|
|
|
|
|
* some random ESSID the user gave us exists or not.
|
|
|
|
|
*
|
|
|
|
|
* WARNING: will blow away any connection the card currently has.
|
|
|
|
|
*
|
|
|
|
|
*/
|
2004-10-29 16:37:43 +00:00
|
|
|
gboolean nm_device_wireless_network_exists (NMDevice *dev, const char *network, const char *key, NMEncKeyType key_type,
|
|
|
|
|
struct ether_addr *ap_addr, gboolean *encrypted)
|
2004-10-13 20:57:23 +00:00
|
|
|
{
|
2004-10-27 19:02:07 +00:00
|
|
|
gboolean success = FALSE;
|
|
|
|
|
struct ether_addr addr;
|
|
|
|
|
NMAccessPoint *ap = NULL;
|
2004-10-13 20:57:23 +00:00
|
|
|
|
|
|
|
|
g_return_val_if_fail (dev != NULL, FALSE);
|
|
|
|
|
g_return_val_if_fail (network != NULL, FALSE);
|
|
|
|
|
g_return_val_if_fail (ap_addr != NULL, FALSE);
|
|
|
|
|
g_return_val_if_fail (strlen (network), FALSE);
|
2004-10-15 17:04:57 +00:00
|
|
|
g_return_val_if_fail (encrypted != NULL, FALSE);
|
2004-10-13 20:57:23 +00:00
|
|
|
|
|
|
|
|
fprintf (stderr, "nm_device_wireless_network_exists () looking for network '%s'...", network);
|
|
|
|
|
|
2004-10-15 17:04:57 +00:00
|
|
|
*encrypted = FALSE;
|
2004-10-13 20:57:23 +00:00
|
|
|
|
|
|
|
|
/* Force the card into Managed/Infrastructure mode */
|
|
|
|
|
nm_device_set_mode_managed (dev);
|
|
|
|
|
|
2004-10-29 19:52:35 +00:00
|
|
|
if ((key_type != NM_ENC_TYPE_UNKNOWN) && key)
|
|
|
|
|
{
|
|
|
|
|
char *hashed_key = NULL;
|
|
|
|
|
switch (key_type)
|
|
|
|
|
{
|
|
|
|
|
case (NM_ENC_TYPE_128_BIT_PASSPHRASE):
|
|
|
|
|
hashed_key = nm_wireless_128bit_key_from_passphrase (key);
|
|
|
|
|
break;
|
|
|
|
|
case (NM_ENC_TYPE_128_BIT_HEX_KEY):
|
|
|
|
|
case (NM_ENC_TYPE_40_BIT_PASSPHRASE):
|
|
|
|
|
case (NM_ENC_TYPE_40_BIT_HEX_KEY):
|
|
|
|
|
case (NM_ENC_TYPE_UNKNOWN):
|
|
|
|
|
hashed_key = g_strdup (key);
|
|
|
|
|
break;
|
|
|
|
|
default:
|
|
|
|
|
break;
|
|
|
|
|
}
|
2004-11-03 07:47:32 +00:00
|
|
|
nm_device_set_enc_key (dev, hashed_key, NM_DEVICE_AUTH_METHOD_SHARED_KEY);
|
2004-10-29 19:52:35 +00:00
|
|
|
g_free (hashed_key);
|
|
|
|
|
}
|
|
|
|
|
else
|
2004-11-03 07:47:32 +00:00
|
|
|
nm_device_set_enc_key (dev, "11111111111111111111111111", NM_DEVICE_AUTH_METHOD_SHARED_KEY);
|
2004-10-13 20:57:23 +00:00
|
|
|
nm_device_set_essid (dev, network);
|
|
|
|
|
|
|
|
|
|
/* Bring the device up and pause to allow card to associate */
|
|
|
|
|
nm_device_bring_up (dev);
|
|
|
|
|
g_usleep (G_USEC_PER_SEC * 2);
|
|
|
|
|
|
|
|
|
|
nm_device_update_link_active (dev, FALSE);
|
|
|
|
|
nm_device_get_ap_address (dev, &addr);
|
|
|
|
|
if (nm_ethernet_address_is_valid (&addr) && nm_device_get_essid (dev))
|
|
|
|
|
{
|
|
|
|
|
nm_device_get_ap_address (dev, ap_addr);
|
2004-10-15 17:04:57 +00:00
|
|
|
success = TRUE;
|
2004-10-23 05:19:28 +00:00
|
|
|
*encrypted = TRUE;
|
2004-10-13 20:57:23 +00:00
|
|
|
}
|
|
|
|
|
else
|
|
|
|
|
{
|
2004-10-23 05:19:28 +00:00
|
|
|
/* Okay, try again in unencrypted mode */
|
2004-10-15 17:04:57 +00:00
|
|
|
|
|
|
|
|
/* Force the card into Managed/Infrastructure mode */
|
|
|
|
|
nm_device_set_mode_managed (dev);
|
|
|
|
|
|
2004-11-03 07:47:32 +00:00
|
|
|
nm_device_set_enc_key (dev, NULL, NM_DEVICE_AUTH_METHOD_NONE);
|
2004-10-15 17:04:57 +00:00
|
|
|
nm_device_set_essid (dev, network);
|
|
|
|
|
|
|
|
|
|
/* Bring the device up and pause to allow card to associate */
|
|
|
|
|
nm_device_bring_up (dev);
|
|
|
|
|
g_usleep (G_USEC_PER_SEC * 2);
|
|
|
|
|
|
|
|
|
|
nm_device_update_link_active (dev, FALSE);
|
|
|
|
|
nm_device_get_ap_address (dev, &addr);
|
|
|
|
|
if (nm_ethernet_address_is_valid (&addr) && nm_device_get_essid (dev))
|
|
|
|
|
{
|
|
|
|
|
nm_device_get_ap_address (dev, ap_addr);
|
|
|
|
|
success = TRUE;
|
2004-10-23 05:19:28 +00:00
|
|
|
*encrypted = FALSE;
|
2004-10-15 17:04:57 +00:00
|
|
|
}
|
2004-10-13 20:57:23 +00:00
|
|
|
}
|
2004-10-15 17:04:57 +00:00
|
|
|
|
2004-10-27 19:02:07 +00:00
|
|
|
/* If by some chance we could connect, but in the wrong encryption mode, return the
|
|
|
|
|
* encryption status of the access point if its in our scan, since that's more accurate.
|
|
|
|
|
*/
|
|
|
|
|
if ((ap = nm_ap_list_get_ap_by_essid (nm_device_ap_list_get (dev), network)))
|
|
|
|
|
*encrypted = nm_ap_get_encrypted (ap);
|
|
|
|
|
|
2004-10-15 17:04:57 +00:00
|
|
|
if (success)
|
2004-10-23 05:19:28 +00:00
|
|
|
fprintf (stderr, " found! (%s)\n", *encrypted ? "encrypted" : "unencrypted");
|
2004-10-15 17:04:57 +00:00
|
|
|
else
|
|
|
|
|
fprintf (stderr, " not found\n");
|
|
|
|
|
|
|
|
|
|
return (success);
|
2004-10-13 20:57:23 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
2004-10-14 20:31:35 +00:00
|
|
|
/*
|
|
|
|
|
* nm_device_find_and_use_essid
|
|
|
|
|
*
|
|
|
|
|
* Given an essid, attempt to associate with that ESSID even if we can't
|
|
|
|
|
* see it in our scan. If we successfully find it, mark that network as
|
|
|
|
|
* our "best" and use it during the next activation.
|
|
|
|
|
*
|
|
|
|
|
* Returns: TRUE on success
|
|
|
|
|
* FALSE on failure
|
|
|
|
|
*/
|
2004-10-29 16:37:43 +00:00
|
|
|
gboolean nm_device_find_and_use_essid (NMDevice *dev, const char *essid, const char *key, NMEncKeyType key_type)
|
2004-10-14 20:31:35 +00:00
|
|
|
{
|
|
|
|
|
struct ether_addr ap_addr;
|
2004-10-15 17:04:57 +00:00
|
|
|
gboolean encrypted = FALSE;
|
2004-10-14 20:31:35 +00:00
|
|
|
NMAccessPoint *ap = NULL;
|
|
|
|
|
gboolean success = FALSE;
|
|
|
|
|
|
|
|
|
|
g_return_val_if_fail (dev != NULL, FALSE);
|
|
|
|
|
g_return_val_if_fail (nm_device_is_wireless (dev), FALSE);
|
|
|
|
|
g_return_val_if_fail (dev->app_data != NULL, FALSE);
|
|
|
|
|
g_return_val_if_fail (essid != NULL, FALSE);
|
|
|
|
|
|
|
|
|
|
syslog (LOG_DEBUG, "Forcing AP '%s'", essid);
|
|
|
|
|
/* If the network exists, make sure it has the correct ESSID set
|
|
|
|
|
* (it might have been a blank ESSID up to this point) and use it.
|
|
|
|
|
*/
|
|
|
|
|
nm_device_deactivate (dev, FALSE);
|
2004-10-29 16:37:43 +00:00
|
|
|
if (nm_device_wireless_network_exists (dev, essid, key, key_type, &ap_addr, &encrypted))
|
2004-10-14 20:31:35 +00:00
|
|
|
{
|
|
|
|
|
if (!(ap = nm_ap_list_get_ap_by_essid (nm_device_ap_list_get (dev), essid)))
|
|
|
|
|
{
|
2004-10-15 17:04:57 +00:00
|
|
|
NMAccessPoint *tmp_ap;
|
2004-10-14 20:31:35 +00:00
|
|
|
|
2004-10-29 16:37:43 +00:00
|
|
|
if (!(ap = nm_device_ap_list_get_ap_by_address (dev, &ap_addr)))
|
2004-10-15 17:04:57 +00:00
|
|
|
{
|
|
|
|
|
/* Okay, the card didn't see it in the scan, Cisco cards sometimes do this.
|
|
|
|
|
* So we make a "fake" access point and add it to the scan list.
|
|
|
|
|
*/
|
|
|
|
|
ap = nm_ap_new ();
|
|
|
|
|
nm_ap_set_encrypted (ap, encrypted);
|
|
|
|
|
nm_ap_set_artificial (ap, TRUE);
|
|
|
|
|
nm_ap_set_address (ap, &ap_addr);
|
|
|
|
|
nm_ap_list_append_ap (nm_device_ap_list_get (dev), ap);
|
2004-10-15 20:05:56 +00:00
|
|
|
nm_ap_unref (ap);
|
2004-10-15 17:04:57 +00:00
|
|
|
}
|
2004-10-14 20:31:35 +00:00
|
|
|
|
2004-10-15 17:04:57 +00:00
|
|
|
/* Now that this AP has an essid, copy over encryption keys and whatnot */
|
|
|
|
|
nm_ap_set_essid (ap, essid);
|
|
|
|
|
if ((tmp_ap = nm_ap_list_get_ap_by_essid (dev->app_data->allowed_ap_list, essid)))
|
|
|
|
|
{
|
2004-10-29 19:52:35 +00:00
|
|
|
if (key_type == NM_ENC_TYPE_UNKNOWN)
|
2004-10-29 16:37:43 +00:00
|
|
|
nm_ap_set_enc_key_source (ap, nm_ap_get_enc_key_source (tmp_ap), nm_ap_get_enc_method (tmp_ap));
|
2004-10-15 17:04:57 +00:00
|
|
|
nm_ap_set_invalid (ap, nm_ap_get_invalid (tmp_ap));
|
|
|
|
|
nm_ap_set_timestamp (ap, nm_ap_get_timestamp (tmp_ap));
|
2004-10-14 20:31:35 +00:00
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* If we found a valid access point, use it */
|
|
|
|
|
if (ap)
|
|
|
|
|
{
|
2004-10-29 19:52:35 +00:00
|
|
|
/* Use the encryption key and type the user sent us if its valid */
|
|
|
|
|
if ((key_type != NM_ENC_TYPE_UNKNOWN) && key)
|
|
|
|
|
nm_ap_set_enc_key_source (ap, key, key_type);
|
2004-10-14 20:31:35 +00:00
|
|
|
nm_device_set_best_ap (dev, ap);
|
|
|
|
|
nm_device_freeze_best_ap (dev);
|
|
|
|
|
nm_device_activation_cancel (dev);
|
|
|
|
|
success = TRUE;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
return (success);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
2004-07-06 01:34:10 +00:00
|
|
|
/*
|
|
|
|
|
* nm_device_do_normal_scan
|
|
|
|
|
*
|
|
|
|
|
* Scan for access points on cards that support wireless scanning.
|
|
|
|
|
*
|
|
|
|
|
*/
|
|
|
|
|
static void nm_device_do_normal_scan (NMDevice *dev)
|
|
|
|
|
{
|
2004-08-05 18:54:29 +00:00
|
|
|
int iwlib_socket;
|
|
|
|
|
NMData *data;
|
2004-07-06 01:34:10 +00:00
|
|
|
|
2004-10-27 19:02:07 +00:00
|
|
|
g_return_if_fail (dev != NULL);
|
2004-07-25 02:40:19 +00:00
|
|
|
g_return_if_fail (dev->app_data != NULL);
|
2004-08-31 Dan Williams <dcbw@redhat.com>
* Remove 'debug' extern global from all files since we now
use syslog()
* src/NetworkManager.[ch]
- Break out routine that get the net.interface property from HAL,
removing that logic from nm_create_device_and_add_to_list()
- (nm_create_device_and_add_to_list): make this a bit more general so
it doesn't do the talking to HAL. Also add arguments to facilitate
the create of test devices.
- (nm_data_mark_state_changed): rename from nm_data_set_state_modified()
- (nm_data_new, main, nm_print_usage): add new argument "--enable-test-devices"
which makes NetworkManager listen for dbus commands to create test
devices, which have no backing hardware. Use when you're on a plane
for example, and/or forgot your wireless card at home. Test devices
_cannot_ be created unless NM is started with --enable-test-devices.
* src/NetworkManagerDbus.[ch]
- New "getLinkActive" method for devices
- New "setLinkActive" method for devices (only works on test devices)
- New "createTestDevice" method on NetworkManager object to create a test
device of a specified type (ie wired, wireless). UDI is created from
scratch, as is the interface name. Only works when NM is started with
--enable-test-devices switch.
- New "removeTestDevice" method on NetworkManager object which removes a
test device. Only works when NM is started with --enable-test-devices
* src/NetworkManagerDevice.[ch]
- Logic to facilitate test devices. Add variables to NMDevice struct to indicate
whether a device is a test device or not, and what its link status is.
- Deal with test devices in most functions. For those that work directly on hardware
special-case test devices.
- (nm_device_new): don't create a test device if test devices weren't enabled on the
command-line.
- (nm_device_update_link_active): split out logic for wired and wireless device link
checking to separate functions to facilitate test device link checking.
- (nm_device_set_enc_key): Since some drivers for wireless cards are daft and
don't make a distinction between System Authentication and Encryption
(namely Cisco aironet), we use Open System auth when setting a WEP key
on the card. We don't deal with Shared Key auth yet.
- (nm_device_activation_worker): split the activation cancel check logic out into
a separate routine nm_device_activation_cancel_if_needed()
- (nm_device_activation_signal_cancel): rename from nm_device_activation_cancel()
- (nm_device_fake_ap_list): Test wireless devices obviously cannot scan, so create
a list of fake access points that they can "see"
- (nm_device_is_test_device): return whether or not a device is a test device
* src/NetworkManagerPolicy.c
- (nm_policy_get_best_device): attempt to deal with wireless network selection,
previously if you "locked"/forced NM to use a wireless device but then
selected a wireless network for NM to use, it would switch to a wired device.
So, if the active device is wireless and it has a "forced" best AP, use it
if the "forced" best AP is still valid
- (nm_state_modification_monitor): deal with NULL best devices, for example
there were no usable network devices, or the last one was removed
* src/backends/NetworkManager*.c
- Deal with test devices, mostly just return success for operations like getting
a DHCP address
* test/nmtestdevices.c
- Test tool to create/remove/link-switch test devices
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@112 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-31 16:09:15 +00:00
|
|
|
|
|
|
|
|
/* Test devices shouldn't get here since we fake the AP list earlier */
|
|
|
|
|
g_return_if_fail (!dev->test_device);
|
|
|
|
|
|
2004-08-05 18:54:29 +00:00
|
|
|
data = (NMData *)dev->app_data;
|
2004-07-06 01:34:10 +00:00
|
|
|
|
|
|
|
|
/* Device must be up before we can scan */
|
|
|
|
|
if (!nm_device_is_up (dev))
|
|
|
|
|
nm_device_bring_up (dev);
|
2004-08-05 18:54:29 +00:00
|
|
|
g_usleep (G_USEC_PER_SEC);
|
2004-07-06 01:34:10 +00:00
|
|
|
|
|
|
|
|
iwlib_socket = iw_sockets_open ();
|
|
|
|
|
if (iwlib_socket >= 0)
|
|
|
|
|
{
|
|
|
|
|
wireless_scan_head scan_results = { NULL, 0 };
|
|
|
|
|
wireless_scan *tmp_ap;
|
|
|
|
|
int err;
|
2004-10-11 21:32:19 +00:00
|
|
|
NMAccessPointList *old_ap_list = NULL;
|
2004-10-27 19:02:07 +00:00
|
|
|
NMAccessPointList *new_scan_list = NULL;
|
|
|
|
|
NMAccessPointList *earliest_scan = NULL;
|
2004-10-13 20:57:23 +00:00
|
|
|
gboolean have_blank_essids = FALSE;
|
2004-10-15 17:04:57 +00:00
|
|
|
NMAPListIter *iter;
|
|
|
|
|
NMAccessPoint *artificial_ap;
|
2004-07-06 01:34:10 +00:00
|
|
|
|
2004-10-21 17:42:14 +00:00
|
|
|
err = iw_scan (iwlib_socket, (char *)nm_device_get_iface (dev), WIRELESS_EXT, &scan_results);
|
2004-07-06 04:45:00 +00:00
|
|
|
if ((err == -1) && (errno == ENODATA))
|
|
|
|
|
{
|
|
|
|
|
/* Card hasn't had time yet to compile full access point list.
|
|
|
|
|
* Give it some more time and scan again. If that doesn't work
|
|
|
|
|
* give up.
|
|
|
|
|
*/
|
|
|
|
|
g_usleep (G_USEC_PER_SEC / 2);
|
2004-10-21 17:42:14 +00:00
|
|
|
err = iw_scan (iwlib_socket, (char *)nm_device_get_iface (dev), WIRELESS_EXT, &scan_results);
|
2004-07-06 04:45:00 +00:00
|
|
|
if (err == -1)
|
|
|
|
|
{
|
|
|
|
|
close (iwlib_socket);
|
|
|
|
|
return;
|
|
|
|
|
}
|
|
|
|
|
}
|
2004-07-06 01:34:10 +00:00
|
|
|
|
2004-10-11 21:32:19 +00:00
|
|
|
/* New list for current scan data */
|
2004-10-27 19:02:07 +00:00
|
|
|
new_scan_list = nm_ap_list_new (NETWORK_TYPE_DEVICE);
|
|
|
|
|
if (!new_scan_list)
|
2004-08-06 18:19:06 +00:00
|
|
|
{
|
|
|
|
|
nm_dispose_scan_results (scan_results.result);
|
|
|
|
|
close (iwlib_socket);
|
|
|
|
|
return;
|
|
|
|
|
}
|
2004-08-02 21:12:40 +00:00
|
|
|
|
2004-10-11 21:32:19 +00:00
|
|
|
/* Shift all previous cached scan results and dispose of the oldest one. */
|
2004-10-27 19:02:07 +00:00
|
|
|
earliest_scan = dev->options.wireless.cached_ap_list3;
|
2004-10-11 21:32:19 +00:00
|
|
|
dev->options.wireless.cached_ap_list3 = dev->options.wireless.cached_ap_list2;
|
|
|
|
|
dev->options.wireless.cached_ap_list2 = dev->options.wireless.cached_ap_list1;
|
2004-10-27 19:02:07 +00:00
|
|
|
dev->options.wireless.cached_ap_list1 = new_scan_list;
|
2004-10-11 21:32:19 +00:00
|
|
|
|
2004-07-06 01:34:10 +00:00
|
|
|
/* Iterate over scan results and pick a "most" preferred access point. */
|
|
|
|
|
tmp_ap = scan_results.result;
|
|
|
|
|
while (tmp_ap)
|
|
|
|
|
{
|
2004-10-13 20:57:23 +00:00
|
|
|
/* We need at least an ESSID or a MAC address for each access point */
|
|
|
|
|
if (tmp_ap->b.has_essid || tmp_ap->has_ap_addr)
|
2004-07-06 01:34:10 +00:00
|
|
|
{
|
|
|
|
|
NMAccessPoint *nm_ap = nm_ap_new ();
|
|
|
|
|
|
|
|
|
|
/* Copy over info from scan to local structure */
|
2004-10-13 20:57:23 +00:00
|
|
|
if (!tmp_ap->b.has_essid || (tmp_ap->b.essid && !strlen (tmp_ap->b.essid)))
|
|
|
|
|
{
|
|
|
|
|
nm_ap_set_essid (nm_ap, NULL);
|
|
|
|
|
have_blank_essids = TRUE;
|
|
|
|
|
}
|
|
|
|
|
else
|
|
|
|
|
nm_ap_set_essid (nm_ap, tmp_ap->b.essid);
|
2004-08-06 18:19:06 +00:00
|
|
|
|
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
|
|
|
if (tmp_ap->b.has_key && (tmp_ap->b.key_flags & IW_ENCODE_DISABLED))
|
|
|
|
|
nm_ap_set_encrypted (nm_ap, FALSE);
|
|
|
|
|
else
|
|
|
|
|
nm_ap_set_encrypted (nm_ap, TRUE);
|
2004-07-06 01:34:10 +00:00
|
|
|
|
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
|
|
|
if (tmp_ap->has_ap_addr)
|
|
|
|
|
nm_ap_set_address (nm_ap, (const struct ether_addr *)(tmp_ap->ap_addr.sa_data));
|
2004-07-06 01:34:10 +00:00
|
|
|
|
2004-09-13 17:43:16 +00:00
|
|
|
nm_ap_set_strength (nm_ap, nm_wireless_qual_to_percent (dev, &(tmp_ap->stats.qual)));
|
2004-07-06 01:34:10 +00:00
|
|
|
|
|
|
|
|
if (tmp_ap->b.has_freq)
|
|
|
|
|
nm_ap_set_freq (nm_ap, tmp_ap->b.freq);
|
|
|
|
|
|
2004-08-05 18:54:29 +00:00
|
|
|
/* Add the AP to the device's AP list */
|
2004-10-11 21:32:19 +00:00
|
|
|
nm_ap_list_append_ap (dev->options.wireless.cached_ap_list1, nm_ap);
|
2004-10-15 20:05:56 +00:00
|
|
|
nm_ap_unref (nm_ap);
|
2004-07-06 01:34:10 +00:00
|
|
|
}
|
|
|
|
|
tmp_ap = tmp_ap->next;
|
|
|
|
|
}
|
|
|
|
|
nm_dispose_scan_results (scan_results.result);
|
|
|
|
|
close (iwlib_socket);
|
2004-08-02 21:12:40 +00:00
|
|
|
|
2004-10-11 21:32:19 +00:00
|
|
|
/* Compose the current access point list for the card based on the past two scans. This
|
|
|
|
|
* is to achieve some stability in the list, since cards don't necessarily return the same
|
|
|
|
|
* access point list each scan even if you are standing in the same place.
|
|
|
|
|
* Once we have the list, copy in any relevant information from our Allowed list.
|
|
|
|
|
*/
|
2004-10-13 20:57:23 +00:00
|
|
|
old_ap_list = nm_device_ap_list_get (dev);
|
2004-10-11 21:32:19 +00:00
|
|
|
dev->options.wireless.ap_list = nm_ap_list_combine (dev->options.wireless.cached_ap_list1, dev->options.wireless.cached_ap_list2);
|
2004-10-13 20:57:23 +00:00
|
|
|
nm_ap_list_copy_properties (nm_device_ap_list_get (dev), dev->app_data->allowed_ap_list);
|
|
|
|
|
|
|
|
|
|
/* If any blank ESSID networks were detected in the current scan, try to match their
|
|
|
|
|
* AP MAC address with existing ones in previous scans, and if we get a match, copy the
|
|
|
|
|
* ESSID over to the newest scan list. This enures that we keep the known ESSID for that
|
|
|
|
|
* base station around as long as possible, which allows nm_device_update_best_ap() to do
|
|
|
|
|
* its job when the user wanted us to connect to a non-broadcasting network.
|
|
|
|
|
*/
|
|
|
|
|
if (have_blank_essids)
|
|
|
|
|
nm_ap_list_copy_essids_by_address (nm_device_ap_list_get (dev), old_ap_list);
|
2004-10-15 17:04:57 +00:00
|
|
|
|
2004-10-27 19:02:07 +00:00
|
|
|
/* Furthermore, if we have any "artificial" access points, ie ones that exist but don't show up in
|
2004-10-15 17:04:57 +00:00
|
|
|
* the scan for some reason, copy those over if we are associated with that access point right now.
|
|
|
|
|
* Some Cisco cards don't report non-ESSID-broadcasting access points in their scans even though
|
|
|
|
|
* the card associates with that AP just fine.
|
|
|
|
|
*/
|
|
|
|
|
if ((iter = nm_ap_list_iter_new (old_ap_list)))
|
|
|
|
|
{
|
|
|
|
|
char *essid = nm_device_get_essid (dev);
|
|
|
|
|
|
|
|
|
|
while (essid && (artificial_ap = nm_ap_list_iter_next (iter)))
|
|
|
|
|
{
|
|
|
|
|
/* Copy over the artificial AP from the old list to the new one if
|
|
|
|
|
* its the AP the card is currently associated with.
|
|
|
|
|
*/
|
|
|
|
|
if ( nm_ap_get_essid (artificial_ap)
|
|
|
|
|
&& !strcmp (essid, nm_ap_get_essid (artificial_ap))
|
|
|
|
|
&& nm_ap_get_artificial (artificial_ap))
|
|
|
|
|
nm_ap_list_append_ap (nm_device_ap_list_get (dev), artificial_ap);
|
|
|
|
|
}
|
|
|
|
|
nm_ap_list_iter_free (iter);
|
|
|
|
|
}
|
2004-10-13 20:57:23 +00:00
|
|
|
nm_ap_list_unref (old_ap_list);
|
2004-10-11 21:32:19 +00:00
|
|
|
|
|
|
|
|
/* Generate the "old" list from the 3rd and 4th oldest scans we've done */
|
2004-10-27 19:02:07 +00:00
|
|
|
old_ap_list = nm_ap_list_combine (dev->options.wireless.cached_ap_list3, earliest_scan);
|
|
|
|
|
|
|
|
|
|
/* Don't need the 4th scan around any more */
|
|
|
|
|
if (earliest_scan)
|
|
|
|
|
nm_ap_list_unref (earliest_scan);
|
2004-10-11 21:32:19 +00:00
|
|
|
|
2004-08-02 21:12:40 +00:00
|
|
|
/* Now do a diff of the old and new networks that we can see, and
|
2004-08-24 Dan Williams <dcbw@redhat.com>
* src/NetworkManagerAP.[ch]
- Add a "enc_method_good" member and accessors to an Access Point
to signal when we've found the correct encryption method
for an access point
- Add a "timestamp" member and accessors, remove "priority" member
and accessors (use timestamps instead)
- Rename "wep_key"->"enc_key"
- (nm_ap_get_enc_key_hashed): new, return the correct mangled key
for a specified encryption method using the access points
source encryption key/passphrase
* src/NetworkManagerAPList.c
- When updating a network with dbus, grab timestamp now instead of
priority
* src/NetworkManagerDBus.[ch]
- Add signal for "DeviceActivating"
- Switch priority->timestamp
* src/NetworkManagerDevice.c
- Change references of "wep_key" -> "enc_key" or "key"
- Signal DeviceActivating when starting activation
- When activating a wireless device, if the access point we are connecting
to is encrypted, and we have a source key, try to generate a mangled
key and use that (ie, generate real WEP key from a passphrase)
- Rework device activation to fallback to other encryption methods if
a previous one didn't work (ie, try mangling a key as a 104-bit passphrase
first, then if that doesn't work fall back to direct hex key).
- (nm_device_update_best_ap): fix a deadlock, and use timestamps instead of
priority. We now prefer the latest access point used, rather than using
a priority scheme
- (nm_device_do_normal_scan): make the encryption method "unknown" on access
points we've just discovered, and merge in correct info from the global
access point lists
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@68 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-25 22:41:12 +00:00
|
|
|
* signal any changes over dbus, but only if we are active device.
|
2004-08-02 21:12:40 +00:00
|
|
|
*/
|
2004-10-13 20:57:23 +00:00
|
|
|
nm_ap_list_diff (dev->app_data, dev, old_ap_list, nm_device_ap_list_get (dev));
|
2004-08-06 18:19:06 +00:00
|
|
|
if (old_ap_list)
|
|
|
|
|
nm_ap_list_unref (old_ap_list);
|
2004-07-06 01:34:10 +00:00
|
|
|
}
|
|
|
|
|
else
|
2004-08-23 19:20:49 +00:00
|
|
|
syslog (LOG_ERR, "nm_device_do_normal_scan() could not get a control socket for the wireless card %s.", nm_device_get_iface (dev) );
|
2004-07-06 01:34:10 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* nm_device_do_pseudo_scan
|
|
|
|
|
*
|
|
|
|
|
* Brute-force the allowed access point list to find one that works, if any.
|
|
|
|
|
*
|
|
|
|
|
* FIXME
|
|
|
|
|
* There's probably a better way to do the non-scanning access point discovery
|
|
|
|
|
* than brute forcing it like this, but that makes the state machine here oh so
|
|
|
|
|
* much more complicated.
|
|
|
|
|
*/
|
|
|
|
|
static void nm_device_do_pseudo_scan (NMDevice *dev)
|
|
|
|
|
{
|
2004-08-05 18:54:29 +00:00
|
|
|
NMAPListIter *iter;
|
|
|
|
|
NMAccessPoint *ap;
|
|
|
|
|
|
2004-07-06 01:34:10 +00:00
|
|
|
g_return_if_fail (dev != NULL);
|
2004-07-25 02:40:19 +00:00
|
|
|
g_return_if_fail (dev->app_data != NULL);
|
2004-07-06 01:34:10 +00:00
|
|
|
|
2004-08-31 Dan Williams <dcbw@redhat.com>
* Remove 'debug' extern global from all files since we now
use syslog()
* src/NetworkManager.[ch]
- Break out routine that get the net.interface property from HAL,
removing that logic from nm_create_device_and_add_to_list()
- (nm_create_device_and_add_to_list): make this a bit more general so
it doesn't do the talking to HAL. Also add arguments to facilitate
the create of test devices.
- (nm_data_mark_state_changed): rename from nm_data_set_state_modified()
- (nm_data_new, main, nm_print_usage): add new argument "--enable-test-devices"
which makes NetworkManager listen for dbus commands to create test
devices, which have no backing hardware. Use when you're on a plane
for example, and/or forgot your wireless card at home. Test devices
_cannot_ be created unless NM is started with --enable-test-devices.
* src/NetworkManagerDbus.[ch]
- New "getLinkActive" method for devices
- New "setLinkActive" method for devices (only works on test devices)
- New "createTestDevice" method on NetworkManager object to create a test
device of a specified type (ie wired, wireless). UDI is created from
scratch, as is the interface name. Only works when NM is started with
--enable-test-devices switch.
- New "removeTestDevice" method on NetworkManager object which removes a
test device. Only works when NM is started with --enable-test-devices
* src/NetworkManagerDevice.[ch]
- Logic to facilitate test devices. Add variables to NMDevice struct to indicate
whether a device is a test device or not, and what its link status is.
- Deal with test devices in most functions. For those that work directly on hardware
special-case test devices.
- (nm_device_new): don't create a test device if test devices weren't enabled on the
command-line.
- (nm_device_update_link_active): split out logic for wired and wireless device link
checking to separate functions to facilitate test device link checking.
- (nm_device_set_enc_key): Since some drivers for wireless cards are daft and
don't make a distinction between System Authentication and Encryption
(namely Cisco aironet), we use Open System auth when setting a WEP key
on the card. We don't deal with Shared Key auth yet.
- (nm_device_activation_worker): split the activation cancel check logic out into
a separate routine nm_device_activation_cancel_if_needed()
- (nm_device_activation_signal_cancel): rename from nm_device_activation_cancel()
- (nm_device_fake_ap_list): Test wireless devices obviously cannot scan, so create
a list of fake access points that they can "see"
- (nm_device_is_test_device): return whether or not a device is a test device
* src/NetworkManagerPolicy.c
- (nm_policy_get_best_device): attempt to deal with wireless network selection,
previously if you "locked"/forced NM to use a wireless device but then
selected a wireless network for NM to use, it would switch to a wired device.
So, if the active device is wireless and it has a "forced" best AP, use it
if the "forced" best AP is still valid
- (nm_state_modification_monitor): deal with NULL best devices, for example
there were no usable network devices, or the last one was removed
* src/backends/NetworkManager*.c
- Deal with test devices, mostly just return success for operations like getting
a DHCP address
* test/nmtestdevices.c
- Test tool to create/remove/link-switch test devices
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@112 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-31 16:09:15 +00:00
|
|
|
/* Test devices shouldn't get here since we fake the AP list earlier */
|
|
|
|
|
g_return_if_fail (!dev->test_device);
|
|
|
|
|
|
2004-07-06 01:34:10 +00:00
|
|
|
nm_device_ref (dev);
|
|
|
|
|
|
2004-10-01 03:05:16 +00:00
|
|
|
if (!(iter = nm_ap_list_iter_new (dev->app_data->allowed_ap_list)))
|
2004-08-05 18:54:29 +00:00
|
|
|
return;
|
|
|
|
|
|
|
|
|
|
nm_device_set_essid (dev, "");
|
|
|
|
|
while ((ap = nm_ap_list_iter_next (iter)))
|
2004-07-06 01:34:10 +00:00
|
|
|
{
|
2004-08-05 18:54:29 +00:00
|
|
|
gboolean valid = FALSE;
|
|
|
|
|
struct ether_addr save_ap_addr;
|
|
|
|
|
struct ether_addr cur_ap_addr;
|
2004-07-06 01:34:10 +00:00
|
|
|
|
2004-08-05 18:54:29 +00:00
|
|
|
if (!nm_device_is_up (dev));
|
|
|
|
|
nm_device_bring_up (dev);
|
2004-07-06 01:34:10 +00:00
|
|
|
|
2004-08-05 18:54:29 +00:00
|
|
|
/* Save the MAC address */
|
|
|
|
|
nm_device_get_ap_address (dev, &save_ap_addr);
|
2004-07-06 01:34:10 +00:00
|
|
|
|
2004-08-05 18:54:29 +00:00
|
|
|
nm_device_set_essid (dev, nm_ap_get_essid (ap));
|
2004-08-24 Dan Williams <dcbw@redhat.com>
* src/NetworkManagerAP.[ch]
- Add a "enc_method_good" member and accessors to an Access Point
to signal when we've found the correct encryption method
for an access point
- Add a "timestamp" member and accessors, remove "priority" member
and accessors (use timestamps instead)
- Rename "wep_key"->"enc_key"
- (nm_ap_get_enc_key_hashed): new, return the correct mangled key
for a specified encryption method using the access points
source encryption key/passphrase
* src/NetworkManagerAPList.c
- When updating a network with dbus, grab timestamp now instead of
priority
* src/NetworkManagerDBus.[ch]
- Add signal for "DeviceActivating"
- Switch priority->timestamp
* src/NetworkManagerDevice.c
- Change references of "wep_key" -> "enc_key" or "key"
- Signal DeviceActivating when starting activation
- When activating a wireless device, if the access point we are connecting
to is encrypted, and we have a source key, try to generate a mangled
key and use that (ie, generate real WEP key from a passphrase)
- Rework device activation to fallback to other encryption methods if
a previous one didn't work (ie, try mangling a key as a 104-bit passphrase
first, then if that doesn't work fall back to direct hex key).
- (nm_device_update_best_ap): fix a deadlock, and use timestamps instead of
priority. We now prefer the latest access point used, rather than using
a priority scheme
- (nm_device_do_normal_scan): make the encryption method "unknown" on access
points we've just discovered, and merge in correct info from the global
access point lists
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@68 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-25 22:41:12 +00:00
|
|
|
if (nm_ap_get_enc_key_source (ap))
|
2004-10-08 05:22:17 +00:00
|
|
|
{
|
|
|
|
|
char *hashed_key = nm_ap_get_enc_key_hashed (ap);
|
2004-11-03 07:47:32 +00:00
|
|
|
nm_device_set_enc_key (dev, hashed_key, NM_DEVICE_AUTH_METHOD_SHARED_KEY);
|
2004-10-08 05:22:17 +00:00
|
|
|
g_free (hashed_key);
|
|
|
|
|
}
|
2004-08-05 18:54:29 +00:00
|
|
|
else
|
2004-11-03 07:47:32 +00:00
|
|
|
nm_device_set_enc_key (dev, NULL, NM_DEVICE_AUTH_METHOD_NONE);
|
2004-07-06 01:34:10 +00:00
|
|
|
|
2004-08-05 18:54:29 +00:00
|
|
|
/* Wait a bit for association */
|
|
|
|
|
g_usleep (G_USEC_PER_SEC);
|
2004-07-06 01:34:10 +00:00
|
|
|
|
2004-08-05 18:54:29 +00:00
|
|
|
/* Do we have a valid MAC address? */
|
|
|
|
|
nm_device_get_ap_address (dev, &cur_ap_addr);
|
|
|
|
|
valid = nm_ethernet_address_is_valid (&cur_ap_addr);
|
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
|
|
|
|
2004-08-05 18:54:29 +00:00
|
|
|
/* If the ap address we had before, and the ap address we
|
|
|
|
|
* have now, are the same, AP is invalid. Certain cards (orinoco)
|
|
|
|
|
* will let the essid change, but the the card won't actually de-associate
|
|
|
|
|
* from the previous access point if it can't associate with the new one
|
|
|
|
|
* (ie signal too weak, etc).
|
|
|
|
|
*/
|
|
|
|
|
if (valid && (memcmp (&save_ap_addr, &cur_ap_addr, sizeof (struct ether_addr)) == 0))
|
|
|
|
|
valid = FALSE;
|
2004-07-06 01:34:10 +00:00
|
|
|
|
2004-08-05 18:54:29 +00:00
|
|
|
if (valid)
|
|
|
|
|
{
|
2004-08-23 19:20:49 +00:00
|
|
|
syslog(LOG_INFO, "%s: setting AP '%s' best", nm_device_get_iface (dev), nm_ap_get_essid (ap));
|
2004-08-05 18:54:29 +00:00
|
|
|
|
|
|
|
|
nm_device_set_best_ap (dev, ap);
|
2004-08-31 Dan Williams <dcbw@redhat.com>
* Remove 'debug' extern global from all files since we now
use syslog()
* src/NetworkManager.[ch]
- Break out routine that get the net.interface property from HAL,
removing that logic from nm_create_device_and_add_to_list()
- (nm_create_device_and_add_to_list): make this a bit more general so
it doesn't do the talking to HAL. Also add arguments to facilitate
the create of test devices.
- (nm_data_mark_state_changed): rename from nm_data_set_state_modified()
- (nm_data_new, main, nm_print_usage): add new argument "--enable-test-devices"
which makes NetworkManager listen for dbus commands to create test
devices, which have no backing hardware. Use when you're on a plane
for example, and/or forgot your wireless card at home. Test devices
_cannot_ be created unless NM is started with --enable-test-devices.
* src/NetworkManagerDbus.[ch]
- New "getLinkActive" method for devices
- New "setLinkActive" method for devices (only works on test devices)
- New "createTestDevice" method on NetworkManager object to create a test
device of a specified type (ie wired, wireless). UDI is created from
scratch, as is the interface name. Only works when NM is started with
--enable-test-devices switch.
- New "removeTestDevice" method on NetworkManager object which removes a
test device. Only works when NM is started with --enable-test-devices
* src/NetworkManagerDevice.[ch]
- Logic to facilitate test devices. Add variables to NMDevice struct to indicate
whether a device is a test device or not, and what its link status is.
- Deal with test devices in most functions. For those that work directly on hardware
special-case test devices.
- (nm_device_new): don't create a test device if test devices weren't enabled on the
command-line.
- (nm_device_update_link_active): split out logic for wired and wireless device link
checking to separate functions to facilitate test device link checking.
- (nm_device_set_enc_key): Since some drivers for wireless cards are daft and
don't make a distinction between System Authentication and Encryption
(namely Cisco aironet), we use Open System auth when setting a WEP key
on the card. We don't deal with Shared Key auth yet.
- (nm_device_activation_worker): split the activation cancel check logic out into
a separate routine nm_device_activation_cancel_if_needed()
- (nm_device_activation_signal_cancel): rename from nm_device_activation_cancel()
- (nm_device_fake_ap_list): Test wireless devices obviously cannot scan, so create
a list of fake access points that they can "see"
- (nm_device_is_test_device): return whether or not a device is a test device
* src/NetworkManagerPolicy.c
- (nm_policy_get_best_device): attempt to deal with wireless network selection,
previously if you "locked"/forced NM to use a wireless device but then
selected a wireless network for NM to use, it would switch to a wired device.
So, if the active device is wireless and it has a "forced" best AP, use it
if the "forced" best AP is still valid
- (nm_state_modification_monitor): deal with NULL best devices, for example
there were no usable network devices, or the last one was removed
* src/backends/NetworkManager*.c
- Deal with test devices, mostly just return success for operations like getting
a DHCP address
* test/nmtestdevices.c
- Test tool to create/remove/link-switch test devices
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@112 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-31 16:09:15 +00:00
|
|
|
nm_data_mark_state_changed (dev->app_data);
|
2004-08-05 18:54:29 +00:00
|
|
|
break;
|
|
|
|
|
}
|
2004-07-06 01:34:10 +00:00
|
|
|
}
|
|
|
|
|
|
2004-08-05 18:54:29 +00:00
|
|
|
nm_ap_list_iter_free (iter);
|
2004-07-06 01:34:10 +00:00
|
|
|
nm_device_unref (dev);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
2004-08-31 Dan Williams <dcbw@redhat.com>
* Remove 'debug' extern global from all files since we now
use syslog()
* src/NetworkManager.[ch]
- Break out routine that get the net.interface property from HAL,
removing that logic from nm_create_device_and_add_to_list()
- (nm_create_device_and_add_to_list): make this a bit more general so
it doesn't do the talking to HAL. Also add arguments to facilitate
the create of test devices.
- (nm_data_mark_state_changed): rename from nm_data_set_state_modified()
- (nm_data_new, main, nm_print_usage): add new argument "--enable-test-devices"
which makes NetworkManager listen for dbus commands to create test
devices, which have no backing hardware. Use when you're on a plane
for example, and/or forgot your wireless card at home. Test devices
_cannot_ be created unless NM is started with --enable-test-devices.
* src/NetworkManagerDbus.[ch]
- New "getLinkActive" method for devices
- New "setLinkActive" method for devices (only works on test devices)
- New "createTestDevice" method on NetworkManager object to create a test
device of a specified type (ie wired, wireless). UDI is created from
scratch, as is the interface name. Only works when NM is started with
--enable-test-devices switch.
- New "removeTestDevice" method on NetworkManager object which removes a
test device. Only works when NM is started with --enable-test-devices
* src/NetworkManagerDevice.[ch]
- Logic to facilitate test devices. Add variables to NMDevice struct to indicate
whether a device is a test device or not, and what its link status is.
- Deal with test devices in most functions. For those that work directly on hardware
special-case test devices.
- (nm_device_new): don't create a test device if test devices weren't enabled on the
command-line.
- (nm_device_update_link_active): split out logic for wired and wireless device link
checking to separate functions to facilitate test device link checking.
- (nm_device_set_enc_key): Since some drivers for wireless cards are daft and
don't make a distinction between System Authentication and Encryption
(namely Cisco aironet), we use Open System auth when setting a WEP key
on the card. We don't deal with Shared Key auth yet.
- (nm_device_activation_worker): split the activation cancel check logic out into
a separate routine nm_device_activation_cancel_if_needed()
- (nm_device_activation_signal_cancel): rename from nm_device_activation_cancel()
- (nm_device_fake_ap_list): Test wireless devices obviously cannot scan, so create
a list of fake access points that they can "see"
- (nm_device_is_test_device): return whether or not a device is a test device
* src/NetworkManagerPolicy.c
- (nm_policy_get_best_device): attempt to deal with wireless network selection,
previously if you "locked"/forced NM to use a wireless device but then
selected a wireless network for NM to use, it would switch to a wired device.
So, if the active device is wireless and it has a "forced" best AP, use it
if the "forced" best AP is still valid
- (nm_state_modification_monitor): deal with NULL best devices, for example
there were no usable network devices, or the last one was removed
* src/backends/NetworkManager*.c
- Deal with test devices, mostly just return success for operations like getting
a DHCP address
* test/nmtestdevices.c
- Test tool to create/remove/link-switch test devices
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@112 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-31 16:09:15 +00:00
|
|
|
/*
|
|
|
|
|
* nm_device_fake_ap_list
|
|
|
|
|
*
|
|
|
|
|
* Fake the access point list, used for test devices.
|
|
|
|
|
*
|
|
|
|
|
*/
|
|
|
|
|
static void nm_device_fake_ap_list (NMDevice *dev)
|
|
|
|
|
{
|
|
|
|
|
#define NUM_FAKE_APS 4
|
|
|
|
|
|
|
|
|
|
int i;
|
|
|
|
|
NMAccessPointList *old_ap_list = nm_device_ap_list_get (dev);
|
|
|
|
|
|
|
|
|
|
char *fake_essids[NUM_FAKE_APS] = { "green", "bay", "packers", "rule" };
|
|
|
|
|
struct ether_addr fake_addrs[NUM_FAKE_APS] = {{{0x70, 0x37, 0x03, 0x70, 0x37, 0x03}},
|
|
|
|
|
{{0x12, 0x34, 0x56, 0x78, 0x90, 0xab}},
|
|
|
|
|
{{0xcd, 0xef, 0x12, 0x34, 0x56, 0x78}},
|
|
|
|
|
{{0x90, 0xab, 0xcd, 0xef, 0x12, 0x34}} };
|
|
|
|
|
guint8 fake_qualities[NUM_FAKE_APS] = { 150, 26, 200, 100 };
|
|
|
|
|
double fake_freqs[NUM_FAKE_APS] = { 3.1416, 4.1416, 5.1415, 6.1415 };
|
|
|
|
|
gboolean fake_enc[NUM_FAKE_APS] = { FALSE, TRUE, FALSE, TRUE };
|
|
|
|
|
|
|
|
|
|
g_return_if_fail (dev != NULL);
|
|
|
|
|
g_return_if_fail (dev->app_data != NULL);
|
|
|
|
|
|
|
|
|
|
dev->options.wireless.ap_list = nm_ap_list_new (NETWORK_TYPE_DEVICE);
|
|
|
|
|
|
|
|
|
|
for (i = 0; i < NUM_FAKE_APS; i++)
|
|
|
|
|
{
|
|
|
|
|
NMAccessPoint *nm_ap = nm_ap_new ();
|
|
|
|
|
NMAccessPoint *list_ap;
|
|
|
|
|
|
|
|
|
|
/* Copy over info from scan to local structure */
|
|
|
|
|
nm_ap_set_essid (nm_ap, fake_essids[i]);
|
|
|
|
|
|
|
|
|
|
if (fake_enc[i])
|
|
|
|
|
nm_ap_set_encrypted (nm_ap, FALSE);
|
|
|
|
|
else
|
|
|
|
|
nm_ap_set_encrypted (nm_ap, TRUE);
|
|
|
|
|
|
|
|
|
|
nm_ap_set_address (nm_ap, (const struct ether_addr *)(&fake_addrs[i]));
|
2004-09-13 17:43:16 +00:00
|
|
|
nm_ap_set_strength (nm_ap, fake_qualities[i]);
|
2004-08-31 Dan Williams <dcbw@redhat.com>
* Remove 'debug' extern global from all files since we now
use syslog()
* src/NetworkManager.[ch]
- Break out routine that get the net.interface property from HAL,
removing that logic from nm_create_device_and_add_to_list()
- (nm_create_device_and_add_to_list): make this a bit more general so
it doesn't do the talking to HAL. Also add arguments to facilitate
the create of test devices.
- (nm_data_mark_state_changed): rename from nm_data_set_state_modified()
- (nm_data_new, main, nm_print_usage): add new argument "--enable-test-devices"
which makes NetworkManager listen for dbus commands to create test
devices, which have no backing hardware. Use when you're on a plane
for example, and/or forgot your wireless card at home. Test devices
_cannot_ be created unless NM is started with --enable-test-devices.
* src/NetworkManagerDbus.[ch]
- New "getLinkActive" method for devices
- New "setLinkActive" method for devices (only works on test devices)
- New "createTestDevice" method on NetworkManager object to create a test
device of a specified type (ie wired, wireless). UDI is created from
scratch, as is the interface name. Only works when NM is started with
--enable-test-devices switch.
- New "removeTestDevice" method on NetworkManager object which removes a
test device. Only works when NM is started with --enable-test-devices
* src/NetworkManagerDevice.[ch]
- Logic to facilitate test devices. Add variables to NMDevice struct to indicate
whether a device is a test device or not, and what its link status is.
- Deal with test devices in most functions. For those that work directly on hardware
special-case test devices.
- (nm_device_new): don't create a test device if test devices weren't enabled on the
command-line.
- (nm_device_update_link_active): split out logic for wired and wireless device link
checking to separate functions to facilitate test device link checking.
- (nm_device_set_enc_key): Since some drivers for wireless cards are daft and
don't make a distinction between System Authentication and Encryption
(namely Cisco aironet), we use Open System auth when setting a WEP key
on the card. We don't deal with Shared Key auth yet.
- (nm_device_activation_worker): split the activation cancel check logic out into
a separate routine nm_device_activation_cancel_if_needed()
- (nm_device_activation_signal_cancel): rename from nm_device_activation_cancel()
- (nm_device_fake_ap_list): Test wireless devices obviously cannot scan, so create
a list of fake access points that they can "see"
- (nm_device_is_test_device): return whether or not a device is a test device
* src/NetworkManagerPolicy.c
- (nm_policy_get_best_device): attempt to deal with wireless network selection,
previously if you "locked"/forced NM to use a wireless device but then
selected a wireless network for NM to use, it would switch to a wired device.
So, if the active device is wireless and it has a "forced" best AP, use it
if the "forced" best AP is still valid
- (nm_state_modification_monitor): deal with NULL best devices, for example
there were no usable network devices, or the last one was removed
* src/backends/NetworkManager*.c
- Deal with test devices, mostly just return success for operations like getting
a DHCP address
* test/nmtestdevices.c
- Test tool to create/remove/link-switch test devices
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@112 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-31 16:09:15 +00:00
|
|
|
nm_ap_set_freq (nm_ap, fake_freqs[i]);
|
|
|
|
|
|
2004-09-08 18:14:42 +00:00
|
|
|
/* Merge settings from wireless networks, mainly Keys */
|
|
|
|
|
if ((list_ap = nm_ap_list_get_ap_by_essid (dev->app_data->allowed_ap_list, nm_ap_get_essid (nm_ap))))
|
2004-08-31 Dan Williams <dcbw@redhat.com>
* Remove 'debug' extern global from all files since we now
use syslog()
* src/NetworkManager.[ch]
- Break out routine that get the net.interface property from HAL,
removing that logic from nm_create_device_and_add_to_list()
- (nm_create_device_and_add_to_list): make this a bit more general so
it doesn't do the talking to HAL. Also add arguments to facilitate
the create of test devices.
- (nm_data_mark_state_changed): rename from nm_data_set_state_modified()
- (nm_data_new, main, nm_print_usage): add new argument "--enable-test-devices"
which makes NetworkManager listen for dbus commands to create test
devices, which have no backing hardware. Use when you're on a plane
for example, and/or forgot your wireless card at home. Test devices
_cannot_ be created unless NM is started with --enable-test-devices.
* src/NetworkManagerDbus.[ch]
- New "getLinkActive" method for devices
- New "setLinkActive" method for devices (only works on test devices)
- New "createTestDevice" method on NetworkManager object to create a test
device of a specified type (ie wired, wireless). UDI is created from
scratch, as is the interface name. Only works when NM is started with
--enable-test-devices switch.
- New "removeTestDevice" method on NetworkManager object which removes a
test device. Only works when NM is started with --enable-test-devices
* src/NetworkManagerDevice.[ch]
- Logic to facilitate test devices. Add variables to NMDevice struct to indicate
whether a device is a test device or not, and what its link status is.
- Deal with test devices in most functions. For those that work directly on hardware
special-case test devices.
- (nm_device_new): don't create a test device if test devices weren't enabled on the
command-line.
- (nm_device_update_link_active): split out logic for wired and wireless device link
checking to separate functions to facilitate test device link checking.
- (nm_device_set_enc_key): Since some drivers for wireless cards are daft and
don't make a distinction between System Authentication and Encryption
(namely Cisco aironet), we use Open System auth when setting a WEP key
on the card. We don't deal with Shared Key auth yet.
- (nm_device_activation_worker): split the activation cancel check logic out into
a separate routine nm_device_activation_cancel_if_needed()
- (nm_device_activation_signal_cancel): rename from nm_device_activation_cancel()
- (nm_device_fake_ap_list): Test wireless devices obviously cannot scan, so create
a list of fake access points that they can "see"
- (nm_device_is_test_device): return whether or not a device is a test device
* src/NetworkManagerPolicy.c
- (nm_policy_get_best_device): attempt to deal with wireless network selection,
previously if you "locked"/forced NM to use a wireless device but then
selected a wireless network for NM to use, it would switch to a wired device.
So, if the active device is wireless and it has a "forced" best AP, use it
if the "forced" best AP is still valid
- (nm_state_modification_monitor): deal with NULL best devices, for example
there were no usable network devices, or the last one was removed
* src/backends/NetworkManager*.c
- Deal with test devices, mostly just return success for operations like getting
a DHCP address
* test/nmtestdevices.c
- Test tool to create/remove/link-switch test devices
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@112 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-31 16:09:15 +00:00
|
|
|
{
|
|
|
|
|
nm_ap_set_timestamp (nm_ap, nm_ap_get_timestamp (list_ap));
|
2004-10-08 05:22:17 +00:00
|
|
|
nm_ap_set_enc_key_source (nm_ap, nm_ap_get_enc_key_source (list_ap), nm_ap_get_enc_method (list_ap));
|
2004-08-31 Dan Williams <dcbw@redhat.com>
* Remove 'debug' extern global from all files since we now
use syslog()
* src/NetworkManager.[ch]
- Break out routine that get the net.interface property from HAL,
removing that logic from nm_create_device_and_add_to_list()
- (nm_create_device_and_add_to_list): make this a bit more general so
it doesn't do the talking to HAL. Also add arguments to facilitate
the create of test devices.
- (nm_data_mark_state_changed): rename from nm_data_set_state_modified()
- (nm_data_new, main, nm_print_usage): add new argument "--enable-test-devices"
which makes NetworkManager listen for dbus commands to create test
devices, which have no backing hardware. Use when you're on a plane
for example, and/or forgot your wireless card at home. Test devices
_cannot_ be created unless NM is started with --enable-test-devices.
* src/NetworkManagerDbus.[ch]
- New "getLinkActive" method for devices
- New "setLinkActive" method for devices (only works on test devices)
- New "createTestDevice" method on NetworkManager object to create a test
device of a specified type (ie wired, wireless). UDI is created from
scratch, as is the interface name. Only works when NM is started with
--enable-test-devices switch.
- New "removeTestDevice" method on NetworkManager object which removes a
test device. Only works when NM is started with --enable-test-devices
* src/NetworkManagerDevice.[ch]
- Logic to facilitate test devices. Add variables to NMDevice struct to indicate
whether a device is a test device or not, and what its link status is.
- Deal with test devices in most functions. For those that work directly on hardware
special-case test devices.
- (nm_device_new): don't create a test device if test devices weren't enabled on the
command-line.
- (nm_device_update_link_active): split out logic for wired and wireless device link
checking to separate functions to facilitate test device link checking.
- (nm_device_set_enc_key): Since some drivers for wireless cards are daft and
don't make a distinction between System Authentication and Encryption
(namely Cisco aironet), we use Open System auth when setting a WEP key
on the card. We don't deal with Shared Key auth yet.
- (nm_device_activation_worker): split the activation cancel check logic out into
a separate routine nm_device_activation_cancel_if_needed()
- (nm_device_activation_signal_cancel): rename from nm_device_activation_cancel()
- (nm_device_fake_ap_list): Test wireless devices obviously cannot scan, so create
a list of fake access points that they can "see"
- (nm_device_is_test_device): return whether or not a device is a test device
* src/NetworkManagerPolicy.c
- (nm_policy_get_best_device): attempt to deal with wireless network selection,
previously if you "locked"/forced NM to use a wireless device but then
selected a wireless network for NM to use, it would switch to a wired device.
So, if the active device is wireless and it has a "forced" best AP, use it
if the "forced" best AP is still valid
- (nm_state_modification_monitor): deal with NULL best devices, for example
there were no usable network devices, or the last one was removed
* src/backends/NetworkManager*.c
- Deal with test devices, mostly just return success for operations like getting
a DHCP address
* test/nmtestdevices.c
- Test tool to create/remove/link-switch test devices
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@112 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-31 16:09:15 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* Add the AP to the device's AP list */
|
|
|
|
|
nm_device_ap_list_add_ap (dev, nm_ap);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (dev == dev->app_data->active_device)
|
|
|
|
|
nm_ap_list_diff (dev->app_data, dev, old_ap_list, nm_device_ap_list_get (dev));
|
|
|
|
|
if (old_ap_list)
|
|
|
|
|
nm_ap_list_unref (old_ap_list);
|
|
|
|
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
2004-07-06 01:34:10 +00:00
|
|
|
/*
|
|
|
|
|
* nm_device_do_wireless_scan
|
|
|
|
|
*
|
|
|
|
|
* Get a list of access points this device can see.
|
|
|
|
|
*
|
|
|
|
|
*/
|
|
|
|
|
void nm_device_do_wireless_scan (NMDevice *dev)
|
|
|
|
|
{
|
2004-08-05 18:54:29 +00:00
|
|
|
g_return_if_fail (dev != NULL);
|
|
|
|
|
g_return_if_fail (dev->app_data != NULL);
|
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
|
|
|
g_return_if_fail (nm_device_is_wireless (dev));
|
2004-07-06 01:34:10 +00:00
|
|
|
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
if (!nm_try_acquire_mutex (dev->options.wireless.scan_mutex, __FUNCTION__))
|
|
|
|
|
return;
|
|
|
|
|
|
2004-08-31 Dan Williams <dcbw@redhat.com>
* Remove 'debug' extern global from all files since we now
use syslog()
* src/NetworkManager.[ch]
- Break out routine that get the net.interface property from HAL,
removing that logic from nm_create_device_and_add_to_list()
- (nm_create_device_and_add_to_list): make this a bit more general so
it doesn't do the talking to HAL. Also add arguments to facilitate
the create of test devices.
- (nm_data_mark_state_changed): rename from nm_data_set_state_modified()
- (nm_data_new, main, nm_print_usage): add new argument "--enable-test-devices"
which makes NetworkManager listen for dbus commands to create test
devices, which have no backing hardware. Use when you're on a plane
for example, and/or forgot your wireless card at home. Test devices
_cannot_ be created unless NM is started with --enable-test-devices.
* src/NetworkManagerDbus.[ch]
- New "getLinkActive" method for devices
- New "setLinkActive" method for devices (only works on test devices)
- New "createTestDevice" method on NetworkManager object to create a test
device of a specified type (ie wired, wireless). UDI is created from
scratch, as is the interface name. Only works when NM is started with
--enable-test-devices switch.
- New "removeTestDevice" method on NetworkManager object which removes a
test device. Only works when NM is started with --enable-test-devices
* src/NetworkManagerDevice.[ch]
- Logic to facilitate test devices. Add variables to NMDevice struct to indicate
whether a device is a test device or not, and what its link status is.
- Deal with test devices in most functions. For those that work directly on hardware
special-case test devices.
- (nm_device_new): don't create a test device if test devices weren't enabled on the
command-line.
- (nm_device_update_link_active): split out logic for wired and wireless device link
checking to separate functions to facilitate test device link checking.
- (nm_device_set_enc_key): Since some drivers for wireless cards are daft and
don't make a distinction between System Authentication and Encryption
(namely Cisco aironet), we use Open System auth when setting a WEP key
on the card. We don't deal with Shared Key auth yet.
- (nm_device_activation_worker): split the activation cancel check logic out into
a separate routine nm_device_activation_cancel_if_needed()
- (nm_device_activation_signal_cancel): rename from nm_device_activation_cancel()
- (nm_device_fake_ap_list): Test wireless devices obviously cannot scan, so create
a list of fake access points that they can "see"
- (nm_device_is_test_device): return whether or not a device is a test device
* src/NetworkManagerPolicy.c
- (nm_policy_get_best_device): attempt to deal with wireless network selection,
previously if you "locked"/forced NM to use a wireless device but then
selected a wireless network for NM to use, it would switch to a wired device.
So, if the active device is wireless and it has a "forced" best AP, use it
if the "forced" best AP is still valid
- (nm_state_modification_monitor): deal with NULL best devices, for example
there were no usable network devices, or the last one was removed
* src/backends/NetworkManager*.c
- Deal with test devices, mostly just return success for operations like getting
a DHCP address
* test/nmtestdevices.c
- Test tool to create/remove/link-switch test devices
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@112 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-31 16:09:15 +00:00
|
|
|
/* Compose a fake list of access points */
|
|
|
|
|
if (dev->test_device)
|
|
|
|
|
{
|
|
|
|
|
nm_device_fake_ap_list (dev);
|
|
|
|
|
nm_unlock_mutex (dev->options.wireless.scan_mutex, __FUNCTION__);
|
|
|
|
|
return;
|
|
|
|
|
}
|
|
|
|
|
|
2004-07-06 01:34:10 +00:00
|
|
|
if (nm_device_get_supports_wireless_scan (dev))
|
|
|
|
|
nm_device_do_normal_scan (dev);
|
|
|
|
|
else
|
|
|
|
|
{
|
|
|
|
|
struct ether_addr ap_addr;
|
|
|
|
|
|
|
|
|
|
/* We can't pseudo-scan without switching APs, therefore
|
|
|
|
|
* if the card has a valid access point and its an allowed
|
|
|
|
|
* access point, don't pseudo-scan for others.
|
|
|
|
|
*/
|
|
|
|
|
nm_device_get_ap_address (dev, &ap_addr);
|
|
|
|
|
if ( !nm_ethernet_address_is_valid (&ap_addr)
|
2004-09-08 18:14:42 +00:00
|
|
|
|| !nm_ap_list_get_ap_by_essid (dev->app_data->allowed_ap_list, nm_device_get_essid (dev))
|
2004-07-06 01:34:10 +00:00
|
|
|
|| !nm_device_get_best_ap (dev))
|
|
|
|
|
{
|
|
|
|
|
nm_device_do_pseudo_scan (dev);
|
|
|
|
|
}
|
|
|
|
|
}
|
2004-08-11 Dan Williams <dcbw@redhat.com>
* info-daemon/NetworkManagerInfo.c:
- (main): clean up Seth's code style
* info-daemon/NetworkManagerInfoDbus.c:
- Use the more aptly-named path/service/interface constants from NetworkManager
- Don't return empty strings ("") as object paths ever, instead return errors
* panel-applet/NMWirelessApplet.c:
- Clean up Seth's code style
* src/NetworkManager.[ch]
- (nm_remove_device_from_list): remove anything having to do with pending_device
- (main, nm_print_usage): change --daemon=[yes|no] -> --no-daemon
* src/NetworkManagerAPList.[ch]
- Move Iter struct right above the iter functions to preserve opacity
- (nm_ap_list_remove_ap): implement
- (nm_ap_list_update_network): deal with errors returned from nm_dbus_get_network_priority(),
remove AP if NetworkManagerInfo doesn't know anything about it
- (nm_ap_list_diff): user NMAPList iterators
- (nm_ap_list_print_members): implement debugging function
* src/NetworkManagerDbus.[ch]
- (nm_dbus_nm_get_active_device): remove anything to do with pending_device
- (nm_dbus_get_user_key_for_network): remove DBusPendingCall stuff (unused),
and move the actual key setting stuff into NetworkManagerDevice.c
- (nm_dbus_get_network_priority): return -1 now on errors
- (nm_dbus_nmi_filter): fix strcmp() error that caused PreferredNetworkUpdate signals to
get lost, and force the active device to update its "best" ap when AP lists change
- (nm_dbus_nm_message_handler): Update conditions for returning "connecting" for a "status"
method call due to pending_device member removal
* src/NetworkManagerDevice.[ch]
- Move NMDevice structure to the top
- Add a wireless scan mutex and a best_ap mutex to the Wireless Options structure
- Remove Pending Action stuff from everywhere
- (nm_device_activation_*): We now "begin" activation and start a thread to do the
activation for us. This thread blocks until all conditions for activation have
been met (ie for wireless devices, we need a valid WEP key and a "best" ap), and
then setup up the interface and runs dhclient. We have to do this because there
is no guaruntee how long dhclient takes, and while we are blocking on it, we cannot
run our main loop and respond to dbus method calls or HAL device removals/inserts
- (nm_device_set_user_key_for_network): Move logic here from NetworkManagerDbus.c so we
can tell nm_device_activation_worker() that we've got a key
- (nm_device_*_best_ap): lock access to best_ap member of Wireless Options structure
- (nm_device_get_path_for_ap): dumb it down so the list doesn't lock against itself when
diffing (AP appear/disappear signal functions make sure the AP is actually in the device's
list)
- (nm_device_update_best_ap): move logic from nm_wireless_is_ap_better() here
* src/NetworkManagerPolicy.c
- Remove anything to do with pending_device
- Adjust device activation to deal with activation-in-worker-thread
* src/NetworkManagerUtils.c
- Clean up locking debugging a bit
* src/NetworkManagerWireless.[ch]
- (nm_wireless_is_ap_better): remove, stick logic in nm_device_update_best_ap(). This function
was badly named and is better as a device function
* panel-applet/.cvsignore: add
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@46 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-11 18:14:02 +00:00
|
|
|
|
|
|
|
|
nm_unlock_mutex (dev->options.wireless.scan_mutex, __FUNCTION__);
|
2004-07-06 01:34:10 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
2004-09-08 18:14:42 +00:00
|
|
|
/* System config data accessors */
|
|
|
|
|
|
|
|
|
|
gboolean nm_device_config_get_use_dhcp (NMDevice *dev)
|
|
|
|
|
{
|
|
|
|
|
g_return_val_if_fail (dev != NULL, 0);
|
|
|
|
|
|
|
|
|
|
return (dev->config_info.use_dhcp);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
void nm_device_config_set_use_dhcp (NMDevice *dev, gboolean use_dhcp)
|
|
|
|
|
{
|
|
|
|
|
g_return_if_fail (dev != NULL);
|
|
|
|
|
|
|
|
|
|
dev->config_info.use_dhcp = use_dhcp;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
guint32 nm_device_config_get_ip4_address (NMDevice *dev)
|
|
|
|
|
{
|
|
|
|
|
g_return_val_if_fail (dev != NULL, 0);
|
|
|
|
|
|
|
|
|
|
return (dev->config_info.ip4_address);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
void nm_device_config_set_ip4_address (NMDevice *dev, guint32 addr)
|
|
|
|
|
{
|
|
|
|
|
g_return_if_fail (dev != NULL);
|
|
|
|
|
|
|
|
|
|
dev->config_info.ip4_address = addr;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
guint32 nm_device_config_get_ip4_gateway (NMDevice *dev)
|
|
|
|
|
{
|
|
|
|
|
g_return_val_if_fail (dev != NULL, 0);
|
|
|
|
|
|
|
|
|
|
return (dev->config_info.ip4_gateway);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
void nm_device_config_set_ip4_gateway (NMDevice *dev, guint32 gateway)
|
|
|
|
|
{
|
|
|
|
|
g_return_if_fail (dev != NULL);
|
|
|
|
|
|
|
|
|
|
dev->config_info.ip4_gateway = gateway;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
guint32 nm_device_config_get_ip4_netmask (NMDevice *dev)
|
|
|
|
|
{
|
|
|
|
|
g_return_val_if_fail (dev != NULL, 0);
|
|
|
|
|
|
|
|
|
|
return (dev->config_info.ip4_netmask);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
void nm_device_config_set_ip4_netmask (NMDevice *dev, guint32 netmask)
|
|
|
|
|
{
|
|
|
|
|
g_return_if_fail (dev != NULL);
|
|
|
|
|
|
|
|
|
|
dev->config_info.ip4_netmask = netmask;
|
|
|
|
|
}
|
2004-10-15 00:06:14 +00:00
|
|
|
guint32 nm_device_config_get_ip4_broadcast (NMDevice *dev)
|
|
|
|
|
{
|
|
|
|
|
g_return_val_if_fail (dev != NULL, 0);
|
|
|
|
|
|
|
|
|
|
return (dev->config_info.ip4_broadcast);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
void nm_device_config_set_ip4_broadcast (NMDevice *dev, guint32 broadcast)
|
|
|
|
|
{
|
|
|
|
|
g_return_if_fail (dev != NULL);
|
|
|
|
|
|
|
|
|
|
dev->config_info.ip4_broadcast = broadcast;
|
|
|
|
|
}
|
2004-09-08 18:14:42 +00:00
|
|
|
|
|
|
|
|
|
2004-06-24 14:18:37 +00:00
|
|
|
/****************************************/
|
|
|
|
|
/* Code ripped from HAL */
|
|
|
|
|
/* minor modifications made for */
|
|
|
|
|
/* integration with NLM */
|
|
|
|
|
/****************************************/
|
|
|
|
|
|
|
|
|
|
/** Read a word from the MII transceiver management registers
|
|
|
|
|
*
|
|
|
|
|
* @param iface Which interface
|
|
|
|
|
* @param location Which register
|
|
|
|
|
* @return Word that is read
|
|
|
|
|
*/
|
|
|
|
|
static guint16 mdio_read (int sockfd, struct ifreq *ifr, int location, gboolean new_ioctl_nums)
|
|
|
|
|
{
|
|
|
|
|
guint16 *data = (guint16 *) &(ifr->ifr_data);
|
|
|
|
|
|
|
|
|
|
data[1] = location;
|
|
|
|
|
if (ioctl (sockfd, new_ioctl_nums ? 0x8948 : SIOCDEVPRIVATE + 1, ifr) < 0)
|
|
|
|
|
{
|
2004-08-23 19:20:49 +00:00
|
|
|
syslog(LOG_ERR, "SIOCGMIIREG on %s failed: %s", ifr->ifr_name, strerror (errno));
|
2004-06-24 14:18:37 +00:00
|
|
|
return -1;
|
|
|
|
|
}
|
|
|
|
|
return data[3];
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static gboolean mii_get_link (NMDevice *dev)
|
|
|
|
|
{
|
|
|
|
|
int sockfd;
|
|
|
|
|
struct ifreq ifr;
|
|
|
|
|
gboolean new_ioctl_nums;
|
|
|
|
|
guint16 status_word;
|
|
|
|
|
gboolean link_active = FALSE;
|
|
|
|
|
|
|
|
|
|
sockfd = socket (AF_INET, SOCK_DGRAM, 0);
|
|
|
|
|
if (sockfd < 0)
|
|
|
|
|
{
|
2004-08-23 19:20:49 +00:00
|
|
|
syslog (LOG_ERR, "cannot open socket on interface %s; errno=%d", nm_device_get_iface (dev), errno);
|
2004-06-24 14:18:37 +00:00
|
|
|
return (FALSE);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
snprintf (ifr.ifr_name, IFNAMSIZ, nm_device_get_iface (dev));
|
|
|
|
|
if (ioctl (sockfd, 0x8947, &ifr) >= 0)
|
|
|
|
|
new_ioctl_nums = TRUE;
|
|
|
|
|
else if (ioctl (sockfd, SIOCDEVPRIVATE, &ifr) >= 0)
|
|
|
|
|
new_ioctl_nums = FALSE;
|
|
|
|
|
else
|
|
|
|
|
{
|
2004-08-23 19:20:49 +00:00
|
|
|
syslog (LOG_ERR, "SIOCGMIIPHY on %s failed: %s", ifr.ifr_name, strerror (errno));
|
2004-06-24 14:18:37 +00:00
|
|
|
close (sockfd);
|
|
|
|
|
return (FALSE);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* Refer to http://www.scyld.com/diag/mii-status.html for
|
|
|
|
|
* the full explanation of the numbers
|
|
|
|
|
*
|
|
|
|
|
* 0x8000 Capable of 100baseT4.
|
|
|
|
|
* 0x7800 Capable of 10/100 HD/FD (most common).
|
|
|
|
|
* 0x0040 Preamble suppression permitted.
|
|
|
|
|
* 0x0020 Autonegotiation complete.
|
|
|
|
|
* 0x0010 Remote fault.
|
|
|
|
|
* 0x0008 Capable of Autonegotiation.
|
|
|
|
|
* 0x0004 Link established ("sticky"* on link failure)
|
|
|
|
|
* 0x0002 Jabber detected ("sticky"* on transmit jabber)
|
|
|
|
|
* 0x0001 Extended MII register exist.
|
|
|
|
|
*
|
|
|
|
|
*/
|
|
|
|
|
|
|
|
|
|
/* We have to read it twice to clear any "sticky" bits */
|
|
|
|
|
status_word = mdio_read (sockfd, &ifr, 1, new_ioctl_nums);
|
|
|
|
|
status_word = mdio_read (sockfd, &ifr, 1, new_ioctl_nums);
|
|
|
|
|
|
|
|
|
|
if ((status_word & 0x0016) == 0x0004)
|
|
|
|
|
link_active = TRUE;
|
|
|
|
|
else
|
|
|
|
|
link_active = FALSE;
|
|
|
|
|
|
|
|
|
|
close (sockfd);
|
|
|
|
|
|
|
|
|
|
return (link_active);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/****************************************/
|
|
|
|
|
/* End Code ripped from HAL */
|
|
|
|
|
/****************************************/
|
2004-08-31 Dan Williams <dcbw@redhat.com>
* Remove 'debug' extern global from all files since we now
use syslog()
* src/NetworkManager.[ch]
- Break out routine that get the net.interface property from HAL,
removing that logic from nm_create_device_and_add_to_list()
- (nm_create_device_and_add_to_list): make this a bit more general so
it doesn't do the talking to HAL. Also add arguments to facilitate
the create of test devices.
- (nm_data_mark_state_changed): rename from nm_data_set_state_modified()
- (nm_data_new, main, nm_print_usage): add new argument "--enable-test-devices"
which makes NetworkManager listen for dbus commands to create test
devices, which have no backing hardware. Use when you're on a plane
for example, and/or forgot your wireless card at home. Test devices
_cannot_ be created unless NM is started with --enable-test-devices.
* src/NetworkManagerDbus.[ch]
- New "getLinkActive" method for devices
- New "setLinkActive" method for devices (only works on test devices)
- New "createTestDevice" method on NetworkManager object to create a test
device of a specified type (ie wired, wireless). UDI is created from
scratch, as is the interface name. Only works when NM is started with
--enable-test-devices switch.
- New "removeTestDevice" method on NetworkManager object which removes a
test device. Only works when NM is started with --enable-test-devices
* src/NetworkManagerDevice.[ch]
- Logic to facilitate test devices. Add variables to NMDevice struct to indicate
whether a device is a test device or not, and what its link status is.
- Deal with test devices in most functions. For those that work directly on hardware
special-case test devices.
- (nm_device_new): don't create a test device if test devices weren't enabled on the
command-line.
- (nm_device_update_link_active): split out logic for wired and wireless device link
checking to separate functions to facilitate test device link checking.
- (nm_device_set_enc_key): Since some drivers for wireless cards are daft and
don't make a distinction between System Authentication and Encryption
(namely Cisco aironet), we use Open System auth when setting a WEP key
on the card. We don't deal with Shared Key auth yet.
- (nm_device_activation_worker): split the activation cancel check logic out into
a separate routine nm_device_activation_cancel_if_needed()
- (nm_device_activation_signal_cancel): rename from nm_device_activation_cancel()
- (nm_device_fake_ap_list): Test wireless devices obviously cannot scan, so create
a list of fake access points that they can "see"
- (nm_device_is_test_device): return whether or not a device is a test device
* src/NetworkManagerPolicy.c
- (nm_policy_get_best_device): attempt to deal with wireless network selection,
previously if you "locked"/forced NM to use a wireless device but then
selected a wireless network for NM to use, it would switch to a wired device.
So, if the active device is wireless and it has a "forced" best AP, use it
if the "forced" best AP is still valid
- (nm_state_modification_monitor): deal with NULL best devices, for example
there were no usable network devices, or the last one was removed
* src/backends/NetworkManager*.c
- Deal with test devices, mostly just return success for operations like getting
a DHCP address
* test/nmtestdevices.c
- Test tool to create/remove/link-switch test devices
git-svn-id: http://svn-archive.gnome.org/svn/NetworkManager/trunk@112 4912f4e0-d625-0410-9fb7-b9a5a253dbdc
2004-08-31 16:09:15 +00:00
|
|
|
|
|
|
|
|
|
|
|
|
|
/****************************************/
|
|
|
|
|
/* Test device routes */
|
|
|
|
|
/****************************************/
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* nm_device_is_test_device
|
|
|
|
|
*
|
|
|
|
|
*/
|
|
|
|
|
gboolean nm_device_is_test_device (NMDevice *dev)
|
|
|
|
|
{
|
|
|
|
|
g_return_val_if_fail (dev != NULL, FALSE);
|
|
|
|
|
|
|
|
|
|
return (dev->test_device);
|
|
|
|
|
}
|