8 years agotrivial: fix misspelling
Dan Williams [Wed, 2 Nov 2011 17:15:40 +0000]
trivial: fix misspelling

8 years agoip6: save autoconf config when starting DHCP
Dan Williams [Wed, 2 Nov 2011 17:09:18 +0000]
ip6: save autoconf config when starting DHCP

Otherwise it just gets lost, and we want it later to combine with
the DHCP config to get the final config that's applied to the interface.

8 years agoip6: print RA flags description
Dan Williams [Wed, 2 Nov 2011 16:54:51 +0000]
ip6: print RA flags description

Helps debugging:


8 years agocore: combine DHCP and RA IPv6 configs when either changes
Dan Williams [Fri, 21 Oct 2011 19:25:35 +0000]
core: combine DHCP and RA IPv6 configs when either changes

Since both RA and DHCP may be run at the same time, we want to
make sure to merge both configs into a final config when either
RA or DHCP changes.  Previously this only happened when RA changed,
but not when DHCP changed or completed.  This caused the config
applied when DHCP completed to not contain the RA-derived address,
which was then removed from the device, which then regressed the
IPv6 RA state, causing a device failure.

Found by Tore Anderson

Oct 18 18:35:00 wrath dhclient[13782]: RCV: Reply message on eth0 from fe80::ca6c:87ff:feab:da5f.
Oct 18 18:35:00 wrath NetworkManager[12390]: <info> (eth0): DHCPv6 state changed nbi -> renew6
Oct 18 18:35:00 wrath NetworkManager[12390]: <debug> [1318955700.642273] [nm-device.c:1582] dhcp6_state_changed(): (eth0): new DHCPv6 client state 7
Oct 18 18:35:00 wrath NetworkManager[12390]: <debug> [1318955700.642282] [nm-dhcp-client.c:1211] ip6_options_to_config(): (eth0): option 'interface'=>'eth0'
Oct 18 18:35:00 wrath NetworkManager[12390]: <debug> [1318955700.642288] [nm-dhcp-client.c:1211] ip6_options_to_config(): (eth0): option 'new_dhcp6_client_id'=>'0:3:0:1:0:30:1b:bc:7f:23'
Oct 18 18:35:00 wrath NetworkManager[12390]: <debug> [1318955700.642294] [nm-dhcp-client.c:1211] ip6_options_to_config(): (eth0): option 'reason'=>'RENEW6'
Oct 18 18:35:00 wrath NetworkManager[12390]: <debug> [1318955700.642300] [nm-dhcp-client.c:1211] ip6_options_to_config(): (eth0): option 'new_dhcp6_name_servers'=>'2001:840:100:: 2001:840:200::'
Oct 18 18:35:00 wrath NetworkManager[12390]: <debug> [1318955700.642305] [nm-dhcp-client.c:1211] ip6_options_to_config(): (eth0): option 'new_dhcp6_server_id'=>'0:3:0:1:c8:6c:87:ab:da:5f'
Oct 18 18:35:00 wrath NetworkManager[12390]: <debug> [1318955700.642311] [nm-dhcp-client.c:1211] ip6_options_to_config(): (eth0): option 'pid'=>'13782'
Oct 18 18:35:00 wrath NetworkManager[12390]: <info> Activation (eth0) Stage 5 of 5 (IPv6 Commit) scheduled...
Oct 18 18:35:00 wrath NetworkManager[12390]: <info> Activation (eth0) Stage 5 of 5 (IPv6 Commit) started...
Oct 18 18:35:00 wrath NetworkManager[12390]: <debug> [1318955700.643641] [nm-system.c:182] sync_addresses(): (eth0): syncing addresses (family 10)
Oct 18 18:35:00 wrath NetworkManager[12390]: <debug> [1318955700.643655] [nm-system.c:235] sync_addresses(): (eth0): removing address '2001:840:3033:20:230:1bff:febc:7f23/64'
Oct 18 18:35:00 wrath NetworkManager[12390]: <debug> [1318955700.643702] [nm-system.c:218] sync_addresses(): (eth0): ignoring IPv6 link-local address
Oct 18 18:35:01 wrath NetworkManager[12390]: <info> Policy set 'Wired connection 1' (eth0) as default for IPv4 routing and DNS.
Oct 18 18:35:01 wrath NetworkManager[12390]: <info> Activation (eth0) Stage 5 of 5 (IPv6 Commit) complete.
Oct 18 18:35:01 wrath NetworkManager[12390]: <debug> [1318955701.656335] [nm-ip6-manager.c:1041] netlink_notification(): netlink notificate type 21
Oct 18 18:35:01 wrath NetworkManager[12390]: <debug> [1318955701.656345] [nm-ip6-manager.c:542] process_addr(): processing netlink new/del address message
Oct 18 18:35:01 wrath NetworkManager[12390]: <debug> [1318955701.656359] [nm-ip6-manager.c:1069] netlink_notification(): (eth0): syncing device with netlink changes
Oct 18 18:35:01 wrath NetworkManager[12390]: <debug> [1318955701.656367] [nm-ip6-manager.c:419] nm_ip6_device_sync_from_netlink(): (eth0): syncing with netlink (ra_flags 0x800000B0) (state/target 'got-address'/'got-address')
Oct 18 18:35:01 wrath NetworkManager[12390]: <debug> [1318955701.656376] [nm-ip6-manager.c:438] nm_ip6_device_sync_from_netlink(): (eth0): netlink address: fe80::230:1bff:febc:7f23
Oct 18 18:35:01 wrath NetworkManager[12390]: <debug> [1318955701.656382] [nm-ip6-manager.c:460] nm_ip6_device_sync_from_netlink(): (eth0): addresses synced (state got-address)
Oct 18 18:35:01 wrath NetworkManager[12390]: <debug> [1318955701.656388] [nm-ip6-manager.c:474] nm_ip6_device_sync_from_netlink(): router advertisement requests parallel DHCPv6
Oct 18 18:35:01 wrath NetworkManager[12390]: <debug> [1318955701.656393] [nm-ip6-manager.c:512] nm_ip6_device_sync_from_netlink(): (eth0): RA-provided address no longer valid
Oct 18 18:35:01 wrath NetworkManager[12390]: <info> (eth0): DHCPv6 client pid 13782 exited with status 0
Oct 18 18:35:01 wrath NetworkManager[12390]: <debug> [1318955701.656448] [nm-device.c:1582] dhcp6_state_changed(): (eth0): new DHCPv6 client state 23
Oct 18 18:35:01 wrath NetworkManager[12390]: <info> (eth0): device state change: activated -> failed (reason 'ip-config-unavailable') [100 120 5]

8 years agocore: decouple initial IP configuration states from device state
Dan Williams [Mon, 10 Oct 2011 04:48:13 +0000]
core: decouple initial IP configuration states from device state

Initial IP configuration can happen during ACTIVATED state if both
v4 and v6 are enabled, but one takes longer than the other.  Thus
various checks throughout the code for IP_CONFIG were incorrect
since they depended on IP configuration only happening during the
IP_CONFIG state.  Fix that by using a separate state for IP config
and using that state for various checks instead of the overall
device state.

8 years agocore: kill stage4 (IP Config Get)
Dan Williams [Mon, 10 Oct 2011 03:50:04 +0000]
core: kill stage4 (IP Config Get)

It was somewhat pointless since the IP config is always known when
stage4 gets scheduled, so why not just pass the config to stage5
immediately?  Also helps consolidate the v4/v6 failure handling
logic and makes the operational flow clearer where both v4 and
v6 are active and proceeding in parallel.

8 years agocore: remove unused 'config' parameter from IP timeout methods
Dan Williams [Fri, 7 Oct 2011 18:08:43 +0000]
core: remove unused 'config' parameter from IP timeout methods

Long ago used to create an IPv4 LL config when DHCP failed, but
we don't do that anymore, and it should be done differently anyway.

8 years agocore: remove unused prototype
Dan Williams [Fri, 7 Oct 2011 17:31:10 +0000]
core: remove unused prototype

8 years agocore: consolidate DHCP failure handling code
Dan Williams [Fri, 7 Oct 2011 17:23:35 +0000]
core: consolidate DHCP failure handling code

8 years agocore: more DHCPv6 code shuffle
Dan Williams [Fri, 7 Oct 2011 17:06:26 +0000]
core: more DHCPv6 code shuffle

No functional changes.

8 years agocore: split DHCPv4 and DHCPv6 state and lease handling code
Dan Williams [Fri, 7 Oct 2011 17:05:51 +0000]
core: split DHCPv4 and DHCPv6 state and lease handling code

8 years agocore: move some more code around
Dan Williams [Fri, 7 Oct 2011 16:53:09 +0000]
core: move some more code around

No functional changes.

8 years agocore: move some code around
Dan Williams [Fri, 7 Oct 2011 16:50:37 +0000]
core: move some code around

No functional changes, but prepare for splitting DHCPv4 and DHCPv6
code apart.

8 years agodevice: allow for applying IPv4 and IPv6 settings in parallel
Mathieu Trudel-Lapierre [Mon, 25 Jul 2011 21:43:48 +0000]
device: allow for applying IPv4 and IPv6 settings in parallel

We don't really need to wait before both IPv4 and IPv6 are established before
applying all the settings to the device. Instead, we can apply each separately
when they are ready, which will bring up the interface sooner.

8 years agowifi: check whether driver supports AP mode (nl80211 only)
Dan Williams [Mon, 10 Oct 2011 06:30:24 +0000]
wifi: check whether driver supports AP mode (nl80211 only)

For future use of wpa_supplicant's lightweight AP mode instead
of using Ad-Hoc for hotspot stuff.

8 years agonl80211: use floats for signal quality calculations
Dan Williams [Thu, 6 Oct 2011 14:34:10 +0000]
nl80211: use floats for signal quality calculations

Since we're working with numbers less than one but greater than
zero, we need floats.  Otherwise stuff just gets clipped.

8 years agobuild: check for nl80211.h too
Dan Williams [Wed, 5 Oct 2011 19:29:01 +0000]
build: check for nl80211.h too

8 years agowifi: simplify nl80211 bitrate reporting
Dan Williams [Wed, 5 Oct 2011 19:06:38 +0000]
wifi: simplify nl80211 bitrate reporting

8 years agowifi: add some logging about nl80211 failures on device setup
Dan Williams [Wed, 5 Oct 2011 15:51:24 +0000]
wifi: add some logging about nl80211 failures on device setup

8 years agotodo: nl80211 work is finished
Dan Williams [Wed, 5 Oct 2011 15:35:49 +0000]
todo: nl80211 work is finished

8 years agowifi: fix some bugs in nl80211 code
Dan Williams [Wed, 5 Oct 2011 15:30:04 +0000]
wifi: fix some bugs in nl80211 code

A forgotten return caused a free of invalid data on success, and
the txrate should be in Kbps.

8 years agowifi: add message when WEXT is used
Dan Williams [Wed, 5 Oct 2011 15:29:45 +0000]
wifi: add message when WEXT is used

8 years agowifi: preliminary nl80211 patch
Johannes Berg [Wed, 5 Oct 2011 11:41:49 +0000]
wifi: preliminary nl80211 patch

Add nl80211 support to wifi utils.

8 years agowifi/mesh: convert OLPC Mesh WEXT usage to wifi utils
Dan Williams [Wed, 5 Oct 2011 15:04:01 +0000]
wifi/mesh: convert OLPC Mesh WEXT usage to wifi utils

8 years agowifi: add 'check_scan' parameter to wifi utils init
Dan Williams [Wed, 5 Oct 2011 14:47:37 +0000]
wifi: add 'check_scan' parameter to wifi utils init

Normal WiFi devices want this (since we dont' support devices that
cannot scan) but OLPC Mesh devices don't.

8 years agowifi: add some mesh bits to wifi utils
Dan Williams [Wed, 5 Oct 2011 14:44:11 +0000]
wifi: add some mesh bits to wifi utils

8 years agowifi: make wireless-helper.h private to src/
Dan Williams [Thu, 29 Sep 2011 22:09:27 +0000]
wifi: make wireless-helper.h private to src/

The less stuff that needs WEXT the better.

8 years agowifi: move WEXT-based WiFi interface detection to wifi-utils
Dan Williams [Thu, 29 Sep 2011 22:06:25 +0000]
wifi: move WEXT-based WiFi interface detection to wifi-utils

8 years agowifi: clean up some usage of wireless-helper.h
Dan Williams [Thu, 29 Sep 2011 21:29:35 +0000]
wifi: clean up some usage of wireless-helper.h

8 years agowifi: split WEXT handling out into its own file
Dan Williams [Thu, 29 Sep 2011 21:27:17 +0000]
wifi: split WEXT handling out into its own file

8 years agowifi: remove dependency on wireless-tools/libiw
Dan Williams [Thu, 29 Sep 2011 20:51:55 +0000]
wifi: remove dependency on wireless-tools/libiw

8 years agowifi: remove some WEXT leftovers
Dan Williams [Thu, 29 Sep 2011 20:46:47 +0000]
wifi: remove some WEXT leftovers

8 years agowifi: split out WEXT handling code to make nl80211 conversion simpler
Dan Williams [Thu, 29 Sep 2011 20:39:13 +0000]
wifi: split out WEXT handling code to make nl80211 conversion simpler

8 years agorelease: bump development version to 0.9.3
Dan Williams [Thu, 10 Nov 2011 03:26:45 +0000]
release: bump development version to 0.9.3

Remember, releases that end in odd numbers are development releases.

8 years agorelease: bump version to 0.9.2
Dan Williams [Wed, 9 Nov 2011 21:28:54 +0000]
release: bump version to 0.9.2

8 years agodhclient: pass DBUS_SYSTEM_BUS_ADDRESS environment variable
David Rothlisberger [Wed, 9 Nov 2011 17:17:04 +0000]
dhclient: pass DBUS_SYSTEM_BUS_ADDRESS environment variable

So that dhclient will, in turn, pass it to the action script
(the action script tries to send a DBus signal to NetworkManager
over the system bus).

Dhclient "execve"s the action script with a hand-constructed environment
that only includes specific variables, plus whatever is passed to
dhclient with "-e".

As far as I know, dhcpcd has no option equivalent to dhclient's "-e".

8 years agorelease: update NEWS
Dan Williams [Mon, 7 Nov 2011 19:33:03 +0000]
release: update NEWS

8 years agoip6: subscribe to route group so we get RTM_NEWROUTE/RTM_DELROUTE (lp:886410)
Dan Williams [Sat, 5 Nov 2011 01:10:00 +0000]
ip6: subscribe to route group so we get RTM_NEWROUTE/RTM_DELROUTE (lp:886410)

Found by Michael Mol; need to subscribe to the group otherwise
we won't get the messages.

8 years agocore: correct shared IP range to be 10.42.x.1 as is documented (bgo #657127)
Jiří Klimeš [Thu, 3 Nov 2011 08:16:56 +0000]
core: correct shared IP range to be 10.42.x.1 as is documented (bgo #657127)

This fixes possible "overflow" of range from 10.42.x.1 to the fourth (on LSB)
or the second (on MSB) octet.

8 years agodns: don't update routing and DNS if no devices were ever managed (lp:862134)
Mathieu Trudel-Lapierre [Wed, 2 Nov 2011 20:28:43 +0000]
dns: don't update routing and DNS if no devices were ever managed (lp:862134)

Avoids blowing away existing routes and resolv.conf if NM never
managed any devices.

8 years agoRevert "dns: don't update routing and DNS if no devices were ever managed"
Dan Williams [Wed, 2 Nov 2011 20:27:24 +0000]
Revert "dns: don't update routing and DNS if no devices were ever managed"

This reverts commit 0051b44a09d67b50b548320233bb8d96d5c8018f.

While fixing up the patch I (dcbw) forgot to include the lp# in
the commit message.

8 years agopolicy: don't update DNS if the hostname hasn't actually changed
Dan Williams [Wed, 2 Nov 2011 20:21:52 +0000]
policy: don't update DNS if the hostname hasn't actually changed

This was causing spurious rewrites of resolv.conf.

8 years agodns: remove unused parameter
Dan Williams [Wed, 2 Nov 2011 20:06:35 +0000]
dns: remove unused parameter

8 years agodns: don't update routing and DNS if no devices were ever managed
Mathieu Trudel-Lapierre [Wed, 2 Nov 2011 20:05:43 +0000]
dns: don't update routing and DNS if no devices were ever managed

Avoids blowing away existing routes and resolv.conf if NM never
managed any devices.

8 years agotrivial: set correct log domain in dhcp client
Thomas Bechtold [Fri, 28 Oct 2011 06:31:12 +0000]
trivial: set correct log domain in dhcp client

just set the correct log domain

8 years agopo: updated Italian translation (bgo #662538)
Milo Casagrande [Tue, 1 Nov 2011 16:17:44 +0000]
po: updated Italian translation (bgo #662538)

8 years agopo: updated French translation (bgo #662851)
Claude Paroz [Tue, 1 Nov 2011 16:11:30 +0000]
po: updated French translation (bgo #662851)

8 years agokeyfile: call correct scheme function when writing certificates (bgo #662836)
Florian Echtler [Tue, 1 Nov 2011 15:32:51 +0000]
keyfile: call correct scheme function when writing certificates (bgo #662836)

8 years agocore: fix building for GLib thread API changes (GLib >= 2.31.0) (bgo #662695)
Jiří Klimeš [Mon, 31 Oct 2011 10:50:28 +0000]
core: fix building for GLib thread API changes (GLib >= 2.31.0) (bgo #662695)

8 years agorelease: bump version to (0.9.2-rc1)
Dan Williams [Fri, 28 Oct 2011 17:53:28 +0000]
release: bump version to (0.9.2-rc1)

8 years agorelease: update NEWS
Dan Williams [Fri, 28 Oct 2011 17:17:03 +0000]
release: update NEWS

8 years agocore: fix deprecated usage of g_value_[get|set]_char() (bgo #662694)
Dan Williams [Fri, 28 Oct 2011 16:30:26 +0000]
core: fix deprecated usage of g_value_[get|set]_char() (bgo #662694)

8 years agowimax: fix link timeout interval
Dan Williams [Fri, 28 Oct 2011 16:27:05 +0000]
wimax: fix link timeout interval

Should be seconds, not milliseconds.

8 years agotrivial: fix GSM setting 'home-only' property nickname
Dan Williams [Thu, 27 Oct 2011 16:24:15 +0000]
trivial: fix GSM setting 'home-only' property nickname

8 years agolibnm-util: add nm_connection_is_type()
Thomas Graf [Tue, 18 Oct 2011 11:48:43 +0000]
libnm-util: add nm_connection_is_type()

Adds a helper nm_connection_is_type(connection, type) which returns TRUE
if a connection is of specified type.

Signed-off-by: Thomas Graf <>

8 years agotrivial: remove stray backtick
Dan Williams [Wed, 26 Oct 2011 19:17:23 +0000]
trivial: remove stray backtick

8 years agotest: add debug helper to make enabling NM, MM, and supplicant debuging easier
Mathieu Trudel-Lapierre [Wed, 26 Oct 2011 19:15:11 +0000]
test: add debug helper to make enabling NM, MM, and supplicant debuging easier

8 years agolibnm-glib: fix allow-none introspection annotations (rh #747302)
Dan Williams [Wed, 26 Oct 2011 15:53:47 +0000]
libnm-glib: fix allow-none introspection annotations (rh #747302)

8 years agolibnm-glib: fix introspection annotations for callbacks and user_data (rh #747302)
Dan Williams [Wed, 26 Oct 2011 15:33:38 +0000]
libnm-glib: fix introspection annotations for callbacks and user_data (rh #747302)

user_data arguments should be annotated with (closure) and callbacks
should be marked (allow none) where it's safe to pass NULL.

8 years agovpn: ensure valid specific object when none was given at activation
Dan Williams [Tue, 25 Oct 2011 17:59:35 +0000]
vpn: ensure valid specific object when none was given at activation

If the caller passed "/" to indicate that NM should choose the default
active connection as the parent of the VPN, make sure we pass that
active connection's object path to the VPN for its specific object
path instead of leaving it "/".

8 years agovpn: print VPN plugin states as strings
Dan Williams [Tue, 25 Oct 2011 16:13:39 +0000]
vpn: print VPN plugin states as strings

8 years agologger: ensure external crash logger buffer is terminated
Dan Williams [Thu, 20 Oct 2011 18:41:21 +0000]
logger: ensure external crash logger buffer is terminated

8 years agodhcp: fix key for dhcp options without "new_" prefix
Jiří Klimeš [Thu, 20 Oct 2011 16:09:42 +0000]
dhcp: fix key for dhcp options without "new_" prefix

There are none at present, but still tmp_key has to be key, not value.

8 years agopolicy: fix setting hostname from DHCP options (rh #719100)
Jiří Klimeš [Thu, 20 Oct 2011 14:51:02 +0000]
policy: fix setting hostname from DHCP options (rh #719100)

This commit fixes a regression introduced by commit

dhclient prefixes options with "new_", however we remove that prefix
before putting options into NMDHCP4Config.

8 years agocore: fix leaked return value of nm_netlink_index_to_iface()
Dan Williams [Tue, 18 Oct 2011 20:28:26 +0000]
core: fix leaked return value of nm_netlink_index_to_iface()

This function now returns allocated strings, so we need to make
sure we free them.

8 years agokeyfile: add another testcase for intlist-like SSID (bgo #660232)
Jiří Klimeš [Tue, 18 Oct 2011 15:01:17 +0000]
keyfile: add another testcase for intlist-like SSID (bgo #660232)

8 years agokeyfile: fix reading/writing SSIDs for all cases (bgo #660232)
Jiří Klimeš [Tue, 18 Oct 2011 14:13:39 +0000]
keyfile: fix reading/writing SSIDs for all cases (bgo #660232)

Even with the previous fix some cases were still undistinguishable. For example,SSID like '11;12;' is both valid an intlist and a string.
So this commit:
- escapes ';' character with '\' when writing, and removes '\' while reading
  This clearly differentiates between intlist x strings.
- changes regex pattern to allow spaces before ';' in intlist format

8 years agokeyfile: fix reading intlist-like SSIDs (ie "101") (lp:874328)
Dan Williams [Fri, 14 Oct 2011 16:20:22 +0000]
keyfile: fix reading intlist-like SSIDs (ie "101") (lp:874328)

Intlists have to end with a ';' since that's how they are written
out, and that's the only way we can actually distinguish between
intlist SSIDs and string SSIDs, really.

8 years agokeyfile: add testcase for intlist-like SSID (lp:874328)
Dan Williams [Fri, 14 Oct 2011 16:17:12 +0000]
keyfile: add testcase for intlist-like SSID (lp:874328)

8 years agolibnm-util: add some more introspection annotations to silence warnings
Dan Winship [Thu, 13 Oct 2011 16:24:04 +0000]
libnm-util: add some more introspection annotations to silence warnings

8 years agorelease: update NEWS
Dan Williams [Wed, 12 Oct 2011 15:02:05 +0000]
release: update NEWS

8 years agopo: updated Croatian translation (bgo #659004)
Tomislav Krznar [Wed, 12 Oct 2011 12:35:14 +0000]
po: updated Croatian translation (bgo #659004)

8 years agopo: updated Galician translation (bgo #657286)
Fran Diéguez [Wed, 12 Oct 2011 12:27:08 +0000]
po: updated Galician translation (bgo #657286)

8 years agosl: updated Slovenian translation (bgo #658392)
Matej Urbančič [Wed, 12 Oct 2011 12:13:22 +0000]
sl: updated Slovenian translation (bgo #658392)

8 years agopo: updated Finish translation (bgo #658823)
Timo Jyrinki [Wed, 12 Oct 2011 12:05:46 +0000]
po: updated Finish translation (bgo #658823)

8 years agopo: updated German translation (bgo #659090)
Christian Kirbach [Wed, 12 Oct 2011 11:58:17 +0000]
po: updated German translation (bgo #659090)

8 years agopo: updated Polish translation (bgo #660391)
Piotr Drąg [Wed, 12 Oct 2011 11:04:09 +0000]
po: updated Polish translation (bgo #660391)

8 years agopo: updated Spanish translation (bgo #660625)
Jorge González [Wed, 12 Oct 2011 10:52:38 +0000]
po: updated Spanish translation (bgo #660625)

8 years agobuild: fix glib version check for g_value_get_schar() (bgo #661271)
Colin Walters [Wed, 12 Oct 2011 08:10:45 +0000]
build: fix glib version check for g_value_get_schar() (bgo #661271)

The function appeared in glib master (which was 2.31.0); it's not
in glib-2-30 branch.

8 years agowifi: clarify supplicant disconnect handling behavior
Dan Williams [Wed, 12 Oct 2011 00:03:36 +0000]
wifi: clarify supplicant disconnect handling behavior

No functional change, but clarify what's supposed to happen when
a disconnect occurs during activation or when the device is activated.

If the connection is an 802.1x or LEAP connection who's password is
NOT_SAVED, then we should request a new password immediately since
the existing one is probably no longer valid (ie, a OTP or token).

Otherwise, start a timer to give the supplicant some time to reconnect.

8 years agoifcfg-rh: fix changing WPA connections to WEP
Jiří Klimeš [Tue, 11 Oct 2011 16:10:26 +0000]
ifcfg-rh: fix changing WPA connections to WEP

When connection was changed with the editor from WPA to WEP, KEY_MGMT was
not cleared. This resulted in infinite loop of ifcfg plugin under some
circumstances ("Available to all users" unchecked, I think).
The re-read connection was regarded as WPA and thus it differed from the
stored one.

8 years agoip6: make sure static routes are flushed for manual configs too
Mathieu Trudel-Lapierre [Mon, 10 Oct 2011 21:36:36 +0000]
ip6: make sure static routes are flushed for manual configs too

Manual configs don't require an IP6Manager object and thus static
routes wouldn't get flushed on device deactivation.

8 years agowifi: fix stop condition of EAP method check while loop
Ludwig Nussel [Fri, 7 Oct 2011 13:21:55 +0000]
wifi: fix stop condition of EAP method check while loop

8 years agobuild: glib 2.30.0 doesn't have g_value_get_schar yet
Ludwig Nussel [Fri, 7 Oct 2011 13:21:56 +0000]
build: glib 2.30.0 doesn't have g_value_get_schar yet

8 years agotest: add missing return values
Thomas Jarosch [Fri, 7 Oct 2011 20:25:34 +0000]
test: add missing return values

Signed-off-by: Thomas Jarosch <>

8 years agowifi: simplify supplicant connect failure handling
Dan Williams [Thu, 6 Oct 2011 23:18:39 +0000]
wifi: simplify supplicant connect failure handling

Make it a bit clearer what's happening and where.  Still need to
do some cleanups to make NM ask for secrets less.

8 years agosupplicant: check supported EAP methods for EAP-FAST
Dan Williams [Thu, 6 Oct 2011 18:31:23 +0000]
supplicant: check supported EAP methods for EAP-FAST

8 years agocli: initialize 'selector' variable to please jhbuild
Larry Reaves [Thu, 6 Oct 2011 10:21:08 +0000]
cli: initialize 'selector' variable to please jhbuild

8 years agodnsmasq: bump DNS cache size to 400
Dan Williams [Wed, 5 Oct 2011 22:28:22 +0000]
dnsmasq: bump DNS cache size to 400

150 seems a bit low these days.

8 years agobuild: add --disable-crashtrace option
David Rothlisberger [Wed, 5 Oct 2011 21:51:29 +0000]
build: add --disable-crashtrace option

Disables GNU backtrace extensions.

8 years agoifnet: fix possible use of uninitialized variable
Thomas Jarosch [Tue, 4 Oct 2011 20:23:38 +0000]
ifnet: fix possible use of uninitialized variable

Detected by cppcheck.

Signed-off-by: Thomas Jarosch <>

8 years agoifcfg-rh: use correct g_free() function
Thomas Jarosch [Tue, 4 Oct 2011 20:22:49 +0000]
ifcfg-rh: use correct g_free() function

Detected by cppcheck.

Signed-off-by: Thomas Jarosch <>

(dcbw: fix more occurances)

8 years agocore: fix builds with libnl2
Dan Williams [Tue, 4 Oct 2011 18:20:11 +0000]
core: fix builds with libnl2

8 years agodocs: make sure libnm-glib.png gets into the dist tarball (bgo #659203)
Dan Williams [Mon, 3 Oct 2011 21:47:26 +0000]
docs: make sure libnm-glib.png gets into the dist tarball (bgo #659203)

8 years agoip6: accept RAs when forwarding
Tore Anderson [Mon, 29 Aug 2011 12:26:40 +0000]
ip6: accept RAs when forwarding

This patch makes NM set the "accept_ra" sysctl to 1 rather than 2. This
causes the kernel to process RAs even if it configured to forward IPv6
traffic on the interface in question.

IPv6 forwarding would likely be enabled on a host running virtualised
operating systems with virtualised network adapters, for example. This
should not prevent NM from successfully activating IPv6 on a
NM-controlled interface configured with IPv6 mode Auto.

8 years agolibnm-util: fix deprecated g_value_get_char() for glib 2.29.90 and later
Dan Williams [Mon, 3 Oct 2011 15:33:12 +0000]
libnm-util: fix deprecated g_value_get_char() for glib 2.29.90 and later

8 years agotrivial: remove unwanted exclamation mark to fix compilation
Jiří Klimeš [Mon, 3 Oct 2011 13:54:25 +0000]
trivial: remove unwanted exclamation mark to fix compilation

8 years agoifcfg-rh: don't use spaces in ifcfg file names (rh #742273)
Jiří Klimeš [Mon, 3 Oct 2011 08:30:50 +0000]
ifcfg-rh: don't use spaces in ifcfg file names (rh #742273)

8 years agocore: improve config error handling
Thomas Bechtold [Fri, 30 Sep 2011 22:22:09 +0000]
core: improve config error handling

  * better error messages
  * fix memory leak in parse_state_file ()
  * create intermediate parent directories as needed for state file

8 years agowifi: add debug logging to new AP merge paths
Dan Williams [Fri, 30 Sep 2011 18:23:52 +0000]
wifi: add debug logging to new AP merge paths