Polkit documentation suggests that <allow_any> applies to all
clients, but that's actually not the case. allow_any,
allow_inactive, and allow_active are evaluated individually based
on whether the user is local and active (allow_active), local and
inactive (allow_inactive), and not local (allow_any). Thus all
three allow options must be specified for any authorization other
than 'no'.
IPSec-based VPNs that use the kernel IPSec functionality don't have tunnel
interfaces. Instead, the IP addresses and routes they create are added
to the parent interface and the kernel handles securing the traffic based
on routing. Modify NM to handle this.
Also modify NM to ensure it doesn't touch routes added externally, and to
preserve various route properties correctly.
A gateway route can only be added, if there exists a device route
for that gateway. Therefore, nm_platform_ip4_route_sync() and
nm_platform_ip6_route_sync() has to add the device routes first,
before adding gateway routes.
Note: usually for all configured addresses, there is also a device
route for the subnet added by the kernel. This means, NM must first
configure the addresses before route_sync, so that these implicit device
routes already exist -- this is however already done correctly.
Signed-off-by: Thomas Haller <thaller@redhat.com>
If a route already exists that matches the network, prefix, gateway,
and metric of a route NM would like to add, don't try to overwrite
the route.
Unlike IP addresses, the kernel doesn't update the details, it
appears to completely replace that route, which might screw up
external tools that added the route originally.
One example of this is IPSec via openswan/libreswan. They add the
routes to the kernel upon connection, and if NM replaces those routes,
IPSec no longer works. While this may be due to kernel bugs or
bad handling of route replacement, there's no reason for NM to touch
routes that it wouldn't materially change anyway.
(yes, we could perhaps use NLM_F_REPLACE in add_kernel_object() only
when we really wanted to replace something, but why ask the kernel
to do the work when it's not required anyway?)
Two issues:
1) routes added by external programs or by users with /sbin/ip should not
be modified, but NetworkManager was always changing those routes' metrics
to match the device priority. This caused the nm_platform_ipX_route_sync()
functions to remove the original, external route (due to mismatched metric)
and re-add the route with the NetworkManager specified metric. Fix that
by not touching routes which came from the kernel.
2) Static routes (from persistent connections) that specified a metric were
getting their metric overwritten with the NetworkManager device priority.
Stop doing that.
Since the platform no longer defaults the metric to 1024, callers of
nm_platform_ip4_route_add() (like NMPolicy's default route handling)
must do that themselves, if they desire this behavior.
Tag addresses and routes with their source. We'll use this later to do
(or not do) operations based on where the item came from.
One thing to note is that when synchronizing items with the kernel, all
items are read as source=KERNEL even when they originally came from
NetworkManager, since the kernel has no way of providing this source
information. This requires the source 'priority', which
nm_ip*_config_add_address() and nm_ip*_config_add_route() must respect
to ensure that NM-owned routes don't have their source overwritten
when merging various IP configs in ip*_config_merge_and_apply().
Also of note is that memcmp() can no longer be used to compare
addresses/routes in nm-platform.c, but this had problems before
anyway with ifindex, so that workaround from nm_platform_ip4_route_sync()
can be removed.
https://bugzilla.gnome.org/show_bug.cgi?id=722843https://bugzilla.redhat.com/show_bug.cgi?id=1005416
VPN services that use the kernel's IPSec stack (like OpenSwan/LibreSWAN)
obviously don't have a tunnel interface, so don't require one. If
those services provide Link-Local IPv6 DNS servers they simply won't
work, but perhaps we can filter those out later or warn about them.
Add properties for things that previously were missing to help out
Qt bindings, among others. Also converts libnm-glib over to use the
new properties, and adds some testcases to ensure libnm-glib works
correctly with them.
https://bugzilla.gnome.org/show_bug.cgi?id=715186
Note that this will cause the nm_device_wifi_get_access_points() to
return hidden-SSID access point objects immediately, which it
previously did not do until added/removed signals were sent by
NetworkManager for a hidden SSID AP. Some clients may not handle
this correctly, but given that they would have crashed when the
first hidden SSID AP was found anyway, they should just be fixed.
With the addition of D-Bus properties for object-array properties in
NetworkManager core, libnm-glib can use these properties instead of
the pseudo-property stuff. However, we need to maintain API and
provide individual added/removed signals for these properties, and
that requires diff-ing the new and old object arrays. Add the
infrastructure for doing that.
The original GetAccessPoints() method call never returned hidden SSID
access points. That's useful though, and the new AccessPoints
property returns all of them too, so add this new method to return
all access points, including hidden SSID ones.
Fix a crash caused by "merge: remove at_console..." when a scan request
comes in via the D-Bus interface. This usage of the device "auth-request"
signal was missed the first time around.
Remove at_console, ensuring that all necessary calls are protected by
PolicyKit authorization (which at_console is redundant with). Allows
sessions that are not necessarily local (like SSH or remote desktop)
to talk to NetworkManager, subject to administrator PolicyKit rules.
at_console permissions as implemented by D-Bus have some problems:
1) it is now fully redundant with PolicyKit and session tracking via
systemd/ConsoleKit
2) it uses a different mechanism than PolicyKit or systemd to determine
sessions and whether the user is on local or not (pam_console)
3) it was never widely implemented across so removing it
harmonizes D-Bus permissions on all supported distros
To that end, remove the at_console section of the D-Bus permissions,
and rely on session-tracking and PolicyKit to ensure operations are
locked down.
No changes are being made to PolicyKit or session-tracking, so any
operations denied by those mechanisms are still denied, and no
permissions are being relaxed. Instead, this should allow remote
users who log in via remote desktop or SSH to inspect network state,
change connection parameters, and start/stop interfaces. Obviously
if you are remote, you should not touch the interface which your
connection is using, but that concern shouldn't prevent all the other
nice stuff that you can do with NM.
https://bugzilla.gnome.org/show_bug.cgi?id=707983https://bugzilla.redhat.com/show_bug.cgi?id=979416
This function returns the number of sessions found, but the return
value wasn't being correctly handled for errors. Also fix the
require_active parameter value to be 100% clear about what NM wants.
While this function only returns the path of the requested connection
(the actual settings are always protected), callers that aren't in
the connection's ACL still probably shouldn't get that, if only to
be pedantic.
Now that the objects get replaced when IP configuration changes
instead of being destroyed and a new one created, they need
PropertiesChanged signals.
(noticed as a result of auditing all exported D-Bus objects)
Given an IPv4 address and prefix for a shared config, figure out
the DHCP address range automatically. To keep things simple we
allow a max of 252 addresses (not including network address,
broadcast address, and the hotspot) no matter what prefix you use,
so if the address is 10.0.10.1, you still only get a range of
10.0.10.2 -> 10.0.10.254.
But we also leave some addresses available above the host address
for static stuff, like we did before. This is done on a sliding
scale from 0 to 8 addresses, where about 1/10th the number of
available addresses are reserved.
https://bugzilla.gnome.org/show_bug.cgi?id=675973
Since IPv6 configuration gets updated every time a router advertisement
comes in, it can lead NM to continuously logging:
NetworkManager: <info> Activation (eth0) Stage 5 of 5 (IPv6 Commit) scheduled...
NetworkManager: <info> Activation (eth0) Stage 5 of 5 (IPv6 Commit) started...
NetworkManager: <info> Activation (eth0) Stage 5 of 5 (IPv6 Commit) complete.
that's annoying. So after the initial configuration is done, make
subsequent IPv6 Commit log messages debug instead of info.
Don't disable IPv6 when we're about to assume a connection that may well
have IPv6 already configured on the interface, which removes all addresses
and routes from the interface and generally Breaks Stuff.
When disconnecting a master device, propagate its NMDeviceStateReason
to the slaves. That way, if the reason is USER_REQUESTED, then the
slaves will be blocked from re-autoconnecting as well.
NMActiveConnection was categorizing all deactivation of master
connections as "failure", and NMActRequest was deactivating all of the
master's slaves with REASON_DEPENDENCY_FAILED no matter what the real
reason was.
In fact, NMActiveConnection only needs to handle the cases where the
master fails before enslaving the device; any failure after that point
will be caught by existing master/slave checks in NMDevice. So update
the code accordingly (and remove the master_failed code from
NMVpnConnection entirely, since no master supports having VPN slaves).
If a master activation failed early (eg, because the virtual device
could not be created), then the slaves were not being notified of the
failure. Fix that.
Add a property to NMDevice that can be used to tell whether the device
is enslaved, and if so what its master is.
This is currently internal-only, but it could be exported later
perhaps.
Virtual devices may be created and destroyed, but we need to keep
their autoconnect state across that. Previously this was handled by
NMManager, but it really belongs with the other autoconnect tracking
in NMPolicy and NMSettingsConnection.
This also fixes a bug where NMPolicy would sometimes decide to
autoactivate a virtual device connection which NMManager would then
have to cancel.
