Daniel Wagner [Mon, 19 Mar 2012 12:50:39 +0000 (13:50 +0100)]
ipconfig: Track ipconfigs enabled/disabled state
Daniel Wagner [Mon, 19 Mar 2012 12:50:38 +0000 (13:50 +0100)]
ipconfig: Do not enable IPv6 ipconfigs when loading the properties
When enabling IPv6 when loading the Service properties, the Service
state machine gets out of sync when AutoConnect is disabled.
At this point we should only load the properties and let the
Service state machine decide when to enable the ipconfigs.
And the good news is that the core handles this already.
Daniel Wagner [Mon, 19 Mar 2012 12:50:37 +0000 (13:50 +0100)]
service: Enable AutoConnect for new ethernet services
We need also to set the Favorite flag to true in order to let
the services to be auto connected.
Daniel Wagner [Mon, 19 Mar 2012 12:50:36 +0000 (13:50 +0100)]
service: Load/store AutoConnect and Favorite properties for ethernet
Daniel Wagner [Mon, 19 Mar 2012 12:50:35 +0000 (13:50 +0100)]
ethernet: Do not connect network on LOWER_UP
Ethernet networks should honor the AutoConnect feature as the rest
of the network types.
Just register the network object at the core when we get LOWER_UP.
Then the core can decide when to connect to the network.
There is no behavior change when AutoConnect is enabled, So if the
user plugs the cable the ethernet network will automatically connect
as it was before.
If AutoConnect is disabled, the core will not connect to the network
automatically.
Marcel Holtmann [Mon, 19 Mar 2012 23:13:09 +0000 (16:13 -0700)]
vpn: Remove wrong include of <glib/ghash.h>
Tomasz Bursztyka [Thu, 15 Mar 2012 08:49:16 +0000 (10:49 +0200)]
wispr: do not take care about http_proxy
Tomasz Bursztyka [Thu, 15 Mar 2012 08:49:15 +0000 (10:49 +0200)]
wispr: do not lookup for proxy if service does not provide any
Jukka Rissanen [Wed, 14 Mar 2012 15:28:59 +0000 (17:28 +0200)]
connection: Try to avoid having two default gateways
It is possible that because the rtnl messages arrive
asynchronously, we might end up having two default gateways
that are both active at the same time. If that happens, we try
to downgrade the one that is not yet active.
Jukka Rissanen [Wed, 14 Mar 2012 15:28:58 +0000 (17:28 +0200)]
connection: Remove obsolete routes for VPN
Jukka Rissanen [Wed, 14 Mar 2012 15:28:57 +0000 (17:28 +0200)]
connection: Check all active gateways in update
There can be multiple active gateways in the system at the same
time so we must check them all when updating gateway info.
Jukka Rissanen [Wed, 14 Mar 2012 15:28:56 +0000 (17:28 +0200)]
connection: Get correct VPN phy link data
VPN could be using wrong phy link data if there are multiple
active links.
Jukka Rissanen [Wed, 14 Mar 2012 15:28:55 +0000 (17:28 +0200)]
connection: Do not set host routes in VPN case
These routes are not needed.
Jukka Rissanen [Wed, 14 Mar 2012 15:28:54 +0000 (17:28 +0200)]
connection: Avoid stale memory access
Get the active gateway pointer only after the gateway hash
has been manipulated by add_gateway(). It is possible that
we are accessing stale pointer otherwise.
Jukka Rissanen [Wed, 14 Mar 2012 15:28:53 +0000 (17:28 +0200)]
service: Check redundant default changed notify
Notify default service changed only if we actually changed
the default service.
Jukka Rissanen [Wed, 14 Mar 2012 15:28:52 +0000 (17:28 +0200)]
service: Allow removing only certain type nameservers
When nameserver are removed by __connman_connection_gateway_remove()
then remove only certain type nameserver (IPv4 or IPv6). This is
needed so that we do not loose IPv4 routes if only IPv6 nameservers
are to be removed. This is needed when there are multiple connected
services.
Jukka Rissanen [Wed, 14 Mar 2012 15:28:51 +0000 (17:28 +0200)]
connection: Remove nameserver routes only once
Jukka Rissanen [Wed, 14 Mar 2012 15:28:50 +0000 (17:28 +0200)]
connection: Trigger service updates only after setting gateways
This is needed so that gateways are set properly when service
triggers online checks.
Jukka Rissanen [Wed, 14 Mar 2012 15:28:49 +0000 (17:28 +0200)]
connection: Fix debugging prints
Jukka Rissanen [Wed, 14 Mar 2012 15:28:48 +0000 (17:28 +0200)]
inet: Fix debugging prints
Marcel Holtmann [Tue, 6 Mar 2012 18:23:11 +0000 (10:23 -0800)]
Release 0.79
Tomasz Bursztyka [Tue, 6 Mar 2012 13:04:45 +0000 (15:04 +0200)]
test: test-session does not exit any more when catching dbus exceptions
Fixes BMC#24964
Tomasz Bursztyka [Tue, 6 Mar 2012 12:45:14 +0000 (14:45 +0200)]
test: Support the possibility to input multiple parameters into simple-agent
Fixes BMC#24965
Tomasz Bursztyka [Tue, 6 Mar 2012 12:45:13 +0000 (14:45 +0200)]
test: Add support of hidden service input into simple-agent
Jussi Kukkonen [Thu, 1 Mar 2012 14:13:44 +0000 (16:13 +0200)]
ofono: Remove network whenever context is removed
Fixes BMC#24942
Jussi Kukkonen [Thu, 1 Mar 2012 14:13:43 +0000 (16:13 +0200)]
ofono: fix modem lookup in cm_context_added()
Marcel Holtmann [Mon, 5 Mar 2012 04:45:54 +0000 (20:45 -0800)]
ipconfig: Fix broken attempt to handle {IPv4,IPv6}.Configuration
Marcel Holtmann [Mon, 5 Mar 2012 04:45:33 +0000 (20:45 -0800)]
service: Fix two small style issues
Marcel Holtmann [Mon, 5 Mar 2012 04:34:19 +0000 (20:34 -0800)]
core: Add extra checks for set property methods
Marcel Holtmann [Mon, 5 Mar 2012 04:33:57 +0000 (20:33 -0800)]
clock: Revert change to accept variant of variant
Daniel Wagner [Thu, 1 Mar 2012 12:37:20 +0000 (13:37 +0100)]
AUTHORS: Metion Tim's contribution
Daniel Wagner [Thu, 1 Mar 2012 12:23:56 +0000 (13:23 +0100)]
AUTHORS: Mention Otavio's contribution
Tim Sander [Thu, 1 Mar 2012 12:35:32 +0000 (13:35 +0100)]
ipconfig: Accept variant types in __connman_ipconfig_set_config()
According the D-Bus API, the Service interface accepts
variant types.
Tim Sander [Thu, 1 Mar 2012 12:29:44 +0000 (13:29 +0100)]
clock: Accept variant types in SetProperty
According the D-Bus API, the Clock interface accepts
variant types.
Patrik Flykt [Wed, 29 Feb 2012 14:25:44 +0000 (16:25 +0200)]
service: Fix ServiceRemoved signals
Removed services hash table need not be consulted when sending
ServicesAdded signals as the service structures have already been
deleted. Also clean up both added and removed hash tables after the
signals have been sent.
Otavio Salvador [Wed, 29 Feb 2012 08:20:24 +0000 (05:20 -0300)]
inet: Ensure data->channel exist before accessing it
Log of segfault:
...
connmand[676]: src/network.c:check_dhcpv6() reply (nil)
connmand[676]: src/network.c:check_dhcpv6() re-send router solicitation 3
connmand[676]: src/inet.c:__connman_inet_ipv6_send_rs()
connmand[676]: Aborting (signal 11) [connmand]
Alok Barsode [Fri, 24 Feb 2012 14:17:59 +0000 (16:17 +0200)]
technology: Enable tech if device is already powered
connman_technology_add_device() calls __connman_device_enable()
but since the device is already enabled, the calls does not propagate
through to __connman_technology_enabled via connman_device_set_powered.
Grant Erickson [Tue, 28 Feb 2012 17:55:53 +0000 (09:55 -0800)]
core: Change coding style for passing NULL to time
Pass NULL rather than 0 when calling time(2) to follow prefered
stylistic convention.
Grant Erickson [Tue, 28 Feb 2012 17:55:52 +0000 (09:55 -0800)]
dhcpv4-client: Change coding style for passing NULL to time
Pass NULL rather than 0 when calling time(2) to follow prefered
stylistic convention.
Grant Erickson [Tue, 28 Feb 2012 16:30:00 +0000 (08:30 -0800)]
gdhcp: set secs in a RFC 951- and 2131-compliant manner
This patch sets the BOOTP secs field in a RFC 951- and 2131-compliant
manner for DHCPv4 DISCOVER and REQUEST/SELECT packets.
Certain DHCP servers, such as that implemented in Mac OS X (< 10.7)
for its "Internet Sharing" feature, refuse to issue a DHCP lease to
clients that have not set a non-zero value in their DISCOVER or
REQUEST packets. In fact, based on http://hints.macworld.com/article.php?
story=
20071223001432304, it's not non-zero but a value greater than four (4)
seconds to allow another "authoritative" DHCP server on the subnet to reply
first.
Side-by-side packet analysis of Mac OS X, iOS, Android, ISC and
Windows clients show that these clients set the BOOTP 'secs' field and
are successfully issued a DHCP lease by Mac OS X. By contrast, a
connman-based client will issue 10 back-to-back DISCOVER packets and
will not be returned a DHCP OFFER from the server.
Daniel Wagner [Tue, 28 Feb 2012 12:40:39 +0000 (13:40 +0100)]
AUTHORS: Mention Mario's contributions
Mario Domenech Goulart [Tue, 28 Feb 2012 12:29:12 +0000 (09:29 -0300)]
ippool: Fix NULL pointer access
Not all allocated blocks info data structure have a pool associated.
Mario Domenech Goulart [Tue, 28 Feb 2012 12:29:11 +0000 (09:29 -0300)]
ippool: Fix typo in error message
Daniel Wagner [Tue, 28 Feb 2012 09:23:31 +0000 (10:23 +0100)]
ofono: Remove obsolete DUN comment
Daniel Wagner [Sun, 26 Feb 2012 12:01:38 +0000 (13:01 +0100)]
test-connman: Use GetServices() instead GetProperties()
GetProperties() does not return the services path anymore. Instead
the GetServices() method should be used.
Daniel Wagner [Sun, 26 Feb 2012 12:01:37 +0000 (13:01 +0100)]
list-services: Use GetServices() instead GetProperties()
GetProperties() does not return the services path anymore. Instead
the GetServices() method should be used.
Daniel Wagner [Sun, 26 Feb 2012 12:01:36 +0000 (13:01 +0100)]
test-connman: Use GetServices() instead GetProperties()
GetProperties() does not return the services path anymore. Instead
the GetServices() method should be used.
Patrik Flykt [Tue, 28 Feb 2012 09:13:34 +0000 (11:13 +0200)]
service: Keep ordering and send all services in ServicesAdded
Keep services ordering and send also already known services
in ServicesAdded signal. Properties are added for new services,
while the dictionary is left empty for already known ones.
When the service is removed, pack the value of "TRUE" to the
removed service hash in order to indicate a removed object path.
The service struct cannot be used, as it is freed immediately
after returning from service_schedule_removed().
Tomasz Bursztyka [Wed, 15 Feb 2012 08:19:50 +0000 (10:19 +0200)]
unit: handling type of connection into session api
Tomasz Bursztyka [Wed, 15 Feb 2012 08:19:49 +0000 (10:19 +0200)]
unit: switching from online to state property into test-session
Tomasz Bursztyka [Wed, 15 Feb 2012 08:19:48 +0000 (10:19 +0200)]
test: modify test-session to handle a new property
Tomasz Bursztyka [Wed, 15 Feb 2012 08:19:47 +0000 (10:19 +0200)]
test: fixing type conversion bug in test-session
Tomasz Bursztyka [Wed, 15 Feb 2012 08:19:46 +0000 (10:19 +0200)]
doc: updated session api documentation
Tomasz Bursztyka [Wed, 15 Feb 2012 08:19:45 +0000 (10:19 +0200)]
session: filtering session state update according to connection type
Tomasz Bursztyka [Wed, 15 Feb 2012 08:19:44 +0000 (10:19 +0200)]
session: introducing a property to select a type on which the state gets up on
Tomasz Bursztyka [Wed, 15 Feb 2012 08:19:43 +0000 (10:19 +0200)]
session: do not notify ipconfig content if it's not in a connected state
Tomasz Bursztyka [Wed, 15 Feb 2012 08:19:42 +0000 (10:19 +0200)]
service: adding a function to know the connected state of an ipconfig
Tomasz Bursztyka [Wed, 15 Feb 2012 08:19:41 +0000 (10:19 +0200)]
session: do not notify an ipconfig change if the session is not connected
Tomasz Bursztyka [Wed, 15 Feb 2012 08:19:40 +0000 (10:19 +0200)]
session: refactor and optimize how notification are raised and when
Tomasz Bursztyka [Wed, 15 Feb 2012 08:19:39 +0000 (10:19 +0200)]
session: changing online property to a state property
Instead of havine online as a boolean, session has now a state which can be
either disconnected, connected or online.
Patrik Flykt [Wed, 22 Feb 2012 09:46:51 +0000 (11:46 +0200)]
AUTHORS: Update email addresses
Marcel Holtmann [Mon, 27 Feb 2012 17:14:57 +0000 (09:14 -0800)]
manager: Fix signature of ServicesRemoved signal
Alok Barsode [Wed, 22 Feb 2012 14:19:13 +0000 (16:19 +0200)]
technology: Fix technology refcounting
__sync_fetch_and_sub() gives the value that had previously been in memory
which gives the older refount.
technology_find() does not refcount the technology.
__connman_technology_update_rfkill() was using technology_get() which
was refcounting the technology. Every technology_get() must be matched
with a corresponding technology_put() for accurate refcounting.
Samuel Ortiz [Fri, 24 Feb 2012 22:06:47 +0000 (23:06 +0100)]
build: Remove find-service and connect-service from test_scripts
Patrik Flykt [Mon, 20 Feb 2012 11:15:40 +0000 (13:15 +0200)]
test: Remove find-service as manager LookupService does not exist
Fixes BMC#24901
Patrik Flykt [Mon, 20 Feb 2012 11:15:08 +0000 (13:15 +0200)]
test: Remove connect-service as manager ConnectService does not exist
Fixes BMC#24900
Patrik Flykt [Mon, 20 Feb 2012 09:43:44 +0000 (11:43 +0200)]
service: Remove 'Services' property from Manager D-Bus API
Remove support functions and 'Services' property from Manager
D-Bus API. Update gateway in all places service was changed.
Patrik Flykt [Mon, 20 Feb 2012 09:43:43 +0000 (11:43 +0200)]
service: Implement ServicesAdded and ServicesRemoved signals
When services are added or removed, collect them into separate
lists in order to send as many as possible in one go and with
then removed ones sent first. Add a wait of 100ms to collect
more services in one batch. When removing a service check
whether it was also on the added list; if so, remove it from
that list.
Patrik Flykt [Mon, 20 Feb 2012 09:43:42 +0000 (11:43 +0200)]
manager: Use helper function for adding arrays of objpath, dict
Use helper function for adding arrays of object path and
dictionary to D-Bus messages.
Patrik Flykt [Mon, 20 Feb 2012 09:43:41 +0000 (11:43 +0200)]
dbus: Function for adding array of object paths and dictionaries
Function for adding an array of object paths and dictionaries to
D-Bus messages.
Daniel Wagner [Mon, 20 Feb 2012 09:00:34 +0000 (10:00 +0100)]
tethering: Fix pool size request
Math is hard, I know.
Fixes BMC#24898
Daniel Wagner [Mon, 20 Feb 2012 10:13:33 +0000 (11:13 +0100)]
ippool: Add error and warning messages
Tomasz Bursztyka [Tue, 14 Feb 2012 09:06:27 +0000 (11:06 +0200)]
wifi: Add support for default background scanning parameters
Tomasz Bursztyka [Tue, 14 Feb 2012 09:06:26 +0000 (11:06 +0200)]
gsupplicant: Add support for background scanning parameters
Alok Barsode [Wed, 15 Feb 2012 11:49:54 +0000 (12:49 +0100)]
manager: Fix TechnologyAdded signal signature
Tomasz Bursztyka [Tue, 14 Feb 2012 08:28:36 +0000 (10:28 +0200)]
test: Fix scan method to work against technology api in test-connman
Tomasz Bursztyka [Tue, 14 Feb 2012 10:05:47 +0000 (12:05 +0200)]
test: Update simple-agent to handle RequestBrowser method
Tomasz Bursztyka [Tue, 14 Feb 2012 10:05:46 +0000 (12:05 +0200)]
doc: Document the right error belonging to browser request on agent api
Tomasz Bursztyka [Tue, 14 Feb 2012 10:05:45 +0000 (12:05 +0200)]
wispr: Provide a callback for agent browser request to handle the result
Tomasz Bursztyka [Tue, 14 Feb 2012 10:05:44 +0000 (12:05 +0200)]
agent: Handle browser request reply
Tomasz Bursztyka [Tue, 14 Feb 2012 10:05:43 +0000 (12:05 +0200)]
wispr: Request a browser action through agent api
Tomasz Bursztyka [Tue, 14 Feb 2012 10:05:42 +0000 (12:05 +0200)]
agent: Add support for requesting browser action
Daniel Wagner [Tue, 14 Feb 2012 12:40:32 +0000 (13:40 +0100)]
service: Set service state idle when creating from provider
When a new service is created from a provider, the service state
should be initialized. We do this also when creating from a
network object.
Santiago Carot-Nemesio [Mon, 13 Feb 2012 17:43:05 +0000 (18:43 +0100)]
agent: Fix capital letter for mandatory requirement
Sometimes, the mandatory attribute is used with capital letter and
other times it isn't. This patch fixes this issue by setting it to
lower case to be consistant with agent-doc in wich it is written
using lower cases.
Grant Erickson [Mon, 13 Feb 2012 17:56:09 +0000 (09:56 -0800)]
dhcp: Request sending a RFC 1533- and 2132-compliant client-id
Some Internet gateways and Wi-Fi access points are unhappy when the
DHCPv4 client-id option (61) is missing and will refuse to issue a
DHCP lease. Adding the client-id option placates these devices.
This patch requests that a RFC 1533- and 2132-compliant DHCP client-id
option (61) be sent with DHCPv4 packets.
Grant Erickson [Mon, 13 Feb 2012 17:56:08 +0000 (09:56 -0800)]
gdhcp: Add RFC 1533- and 2132-compliant client-id option
This patch adds a function to add a RFC 1533- and 2132-compliant DHCP
client-id option (61) to sent DHCPv4 packets.
Grant Erickson [Mon, 13 Feb 2012 17:56:07 +0000 (09:56 -0800)]
gdhcp: Refactor alloc_dhcp_option to accept string or raw data.
Support specifying NULL-terminated string or raw data for allocated
DHCP options.
Daniel Wagner [Thu, 9 Feb 2012 10:02:17 +0000 (11:02 +0100)]
test-ippool: Test overlapping ranges
Daniel Wagner [Thu, 9 Feb 2012 10:00:35 +0000 (11:00 +0100)]
test-ippool: Rename basic2 to exhaust0
Daniel Wagner [Thu, 9 Feb 2012 09:59:11 +0000 (10:59 +0100)]
test-ippool: Reinit ippool for each test
Daniel Wagner [Thu, 9 Feb 2012 09:57:37 +0000 (10:57 +0100)]
test-ippool: Split basic0 test
Do not test two things in the same function.
Daniel Wagner [Thu, 9 Feb 2012 09:50:35 +0000 (10:50 +0100)]
ippool: Fix collision detection altorithm
Track only private address ranges in a list. If the first IP is
assigned for a range check if a pool collides. For this we need
to check all entries in the list if the new IP is the range of an
allocated pool.
This can be made faster with the right data structure and
algorithm (e.g. segment overlap detection algorithms).
Daniel Wagner [Thu, 9 Feb 2012 09:45:15 +0000 (10:45 +0100)]
ippool: Fix 16 bit block iterator
Daniel Wagner [Mon, 6 Feb 2012 15:49:34 +0000 (16:49 +0100)]
ippool: Add prefixlen argument to __connman_ippool_new/deladdr()
Daniel Wagner [Mon, 13 Feb 2012 14:51:42 +0000 (15:51 +0100)]
test-nat: Add main table tests
This test makes use of the iptables check command.
Daniel Wagner [Mon, 13 Feb 2012 14:51:41 +0000 (15:51 +0100)]
test-nat: Commit table changes
Daniel Wagner [Mon, 13 Feb 2012 14:51:40 +0000 (15:51 +0100)]
nat: Fix typo in iptables rule
Daniel Wagner [Mon, 13 Feb 2012 14:51:39 +0000 (15:51 +0100)]
bridge: Fix __connman_bridge_enable() arguments ordering
Daniel Wagner [Mon, 13 Feb 2012 14:51:38 +0000 (15:51 +0100)]
iptables: Mask address when parsing ip/prefixlen
Netfilter likes to have the address properly masked.
Daniel Wagner [Mon, 13 Feb 2012 14:51:37 +0000 (15:51 +0100)]
iptables-test: Mask address when parsing ip/prefixlen
Netfilter likes to have the address properly masked.