Heavily modify Inaky's Intel WiMAX SDK glue (originally from connman)
to be more generic and more thread-safe, and suitable for use with
NetworkManager instead of rolling our own client code. Rewrite the
NMDeviceWimax code to mostly work.
Still to be done: actual connection logic, DHCP handling, spawning
wimaxd if it's not started yet
We're just going to quit here anyway, so we don't need to
make sure the property filter is unregistered. Fixes a
crash on exit if D-Bus has quit before NM has.
It's a leftover from when /etc/hosts got fixed up, and since that
isn't being done anymore, there's no reason to touch the hostname
on exit. It will already have been updated in response to device
deactivation (if the hostname was DHCP-provided for example) so
it shouldn't need to be touched on exit.
This also removes a potential crash on shutdown when systemd or
another startup manager kills D-Bus before NetworkManager, which
made the dispatcher code angry (which got executed on hostname
change).
Handling of /etc/hosts is highly site- and admin- specific in
many more complex cases, and it's exceedingly hard and error-
prone for NetworkManager to handle all those cases. So remove
this functionality entirely. That's not a big loss, as it
turns out there's a much more elegant solution.
The only requirement is that the machine's hostname map back
to an IP address owned by the machine. That requirement can
be satisifed by nss-myhostname or even possibly the distro's
installer. If the user does not want nss-myhostname then it
can be uninstalled. Distros should use a "recommends" feature
in their packaging system so that the NetworkManager package
does *not* have a hard requirement on nss-myhostname. Thus
everyone is happy; things Just Work when nss-myhostname is
installed, but more advanced users can uninstall it and
customize /etc/hosts as they wish.
Another alternative is a dispatcher script that listents for
the 'hostname' event, and updates /etc/hosts according to the
administrator's preference.
When split DNS is used for a local caching nameserver, make sure
that reverse DNS queries for hosts within the VPN tunnel are directed
to the VPN's nameservers, not to the public upstream nameservers.
RFC6101 adds the DNS Search List option to router advertisements. This
allows stateless configuration of suffixes to try when doing DNS lookups.
Make sure we catch these when provided by the kernel and reconfigure
things appropriately.
NOTE: this commit depends on a kernel patch:
http://marc.info/?l=linux-netdev&m=129216173321352&w=2
RFC5006/RFC6106 specifies fairly clearly how to handle multiple RDDNS options.
Unfortunately the previous code didn't deal with this and hence would
misbehave in all but the simplest setups. The new code should be fully
compliant with the following exceptions:
- Router lifetime not respected
- No "sufficient number" management.
- DHCPv6 servers might not be prioritised over RDDNS ones.
Filter registered agents for each secrets request to ensure that the
connection for which secrets are requested is visible to that agent,
and add that agent to the queue. Ask each agent in the queue until
one returns usable secrets. Ensure that if new agents register
or existing agents quit during the secrets request, that the queue
is updated accordingly, and ensure that an agent that's already
been asked for secrets, unregisters, and re-registers before the
secrets request is comple, isn't asked for secrets twice.
Instead of a bizare mechanism of signals back to the manager
object that used to be required because of the user/system settings
split, let each place that needs secrets request those secrets
itself. This flattens the secrets request process a ton and
the code flow significantly.
Previously the get secrets flow was something like this:
nm_act_request_get_secrets ()
nm_secrets_provider_interface_get_secrets ()
emits manager-get-secrets signal
provider_get_secerts ()
system_get_secrets ()
system_get_secrets_idle_cb ()
nm_sysconfig_connection_get_secrets ()
system_get_secrets_reply_cb ()
nm_secrets_provider_interface_get_secrets_result ()
signal failure or success
now instead we do something like this:
nm_agent_manager_get_secrets ()
nm_agent_manager_get_secrets ()
request_start_secrets ()
nm_sysconfig_connection_get_secrets ()
return failure or success to callback
We only want to prevent regression to > READY after READY has
been reached, since the interface state will track the supplicant
connection state which legitimately jumps around.
This branch now requires wpa_supplicant 0.7 or later with the new
D-Bus interface enabled. It also prefers the "nl80211" supplicant
driver over the WEXT one, but will ask the supplicant to fall back
to the WEXT driver if the device in question does not support the
kernel's nl80211/cfg80211 stack.
If the supplicant cannot be service activated, wait until it shows up
on the bus instead of sitting around doing nothing. This fixes a small
regression introduced when the _READY state was added to the supplicant
interface object.
New connections should not be pushed out in the Updated signal
because signals cannot be restricted to particular clients, and
some clients may not have permission to view the connection.
Upon receiving the Updated signal, clients should re-read the
connection using GetSettings to ensure that the client still
has permissions to view the connection, and to get the updated
settings.
Don't bother trying to initialize the user settings proxy when it
isn't even active. Would trigger a warning to syslog if PolicyKit
permissions changed via changes to /usr/share/polkit-1/actions/.
By the time we get the udev device removal notification we may not
be able to read properties, since the device has already been
removed from sysfs. That means we can't get the ifindex, so we need
to fall back to the interface name. But we always want to prefer
the ifindex since that will never change as long as the device is
connected, unlike the interface name.
