Since commit 0922a17738 ("manager: avoid that auto-activations
preempt user activations") the manager doesn't allow a internal
activation to disconnect the same connection already active on the
device. Thus, during a rollback we must ensure that the device is
deactivated before.
Fixes: 0922a17738
Revert the change to initialize the GDBusObjectManager
synchronously. I still maintain that GDBusObjectManager's
async initialization is severely broken on it's own, but this
change broke
$ systemctl restart NetworkManager
$ nmcli connection up "$NAME"
leaving nmcli hanging, as libnm fails to process some D-Bus signals.
I think that hints to another bug, but apparently the previous
version was less broken. So revert for now.
https://bugzilla.redhat.com/show_bug.cgi?id=1450075
Strangely, this breaks
systemctl restart NetworkManager
nmcli connection up "$NAME"
It seems that with this change, libnm misses some events from D-Bus.
It looks like there is something seriously broken. Before fixing it,
revert the previous state.
https://bugzilla.redhat.com/show_bug.cgi?id=1450075
This reverts commit 529d620a59.
On ppc archtecture the "nm_bt_vtable_network_server" symbol in the small
objects section instead of .bss, represencted by a "S" letter. Also
include "G" which is an equivalent thing for initialized data. We don't
seem to have such objects at the moment, but when we do it could result
in a nasty surprise.
For consistency, never return an empty array @values.
If we have an empty array, instead return NULL.
Also fixes commit afac7621a "clients: return NULL array on
auto-completion failure", which claims that readline crashes
with empty strv arrays.
Fixes: afac7621ae
$ nmcli --complete-args connection import type non-existing-<TAB>
Leads to a double-free of out_to_free, as we call g_free(v) in
nm_meta_abstract_info_complete().
Also fix a memleak when skipping over non-matching values.
Fixes: afac7621ae
Commit df0dc912cc ("8021x: don't request secrets if they are empty
and system owned") changed need_private_key_password() to return FALSE
when flags are NONE. This broke authentication using an encrypted
private key because after this the key password is never added to the
applied connection.
Don't require a password with NONE flags only for the PKCS11 scheme.
Fixes: df0dc912cc
The default value for miimon, when missing in the setting, is 0 if
arp_interval is != 0, and 100 otherwise. So, when generating a
connection, let's ignore miimon=0 (which means that miimon is
disabled) and accept any other value. Adding miimon=100 does not cause
any harm to the connection assumption.
While at it, slightly improve the code: ignore_if_zero() is not useful
for 'updelay','downdelay','arp_interval' because zero is their default
value, so introduce a new function that checks if the value is the
default (and specially handles 'miimon').
Reported-by: Taketo Kabe <rkabe@vega.pgw.jp>
https://bugzilla.redhat.com/show_bug.cgi?id=1463077
Since properties are asked only when the connection has the related
setting, ensure that the connection type is set early so that the base
type gets added to the connection before evaluating other settings.
After NMSettingConnection properties, ask properties for the base
setting and then all other settings.
The order matters for the 'nmcli connection show' output and for the
interactive mode of nmcli. Users should not rely on the order in both
cases, but since we have an extensive test suite for the interactive
mode, restore the order as it was in 1.8.
For master devices, instead of ignoring loss of carrier entirely,
handle it.
First of all, master devices are now by default ignore-carrier=yes.
That means, without explict user configuration in NetworkManager.conf,
the previous behavior in carrier_changed() does not change.
If the user decides to configure the master device like
[device-with-carrier]
match-device=type:bond,type:bridge,type:team
ignore-carrier=no
then, master device will disconnect on carrier loss like
regular devices.
https://github.com/NetworkManager/NetworkManager/pull/18
Co-authored-by: Thomas Haller <thaller@redhat.com>
Previously, master device types like bridge, bond, and team
would overwrite is_available() and check_connection_available()
and always return TRUE.
The device already expresses via nm_device_is_master() that it
is of a master kind. Refactor the code, so, instead of having these
device types overwrite is_available() and check_connection_available(),
let the parents implementation react on nm_device_is_master().
There is no change in behavior at all. Instead, the knowledge how to
treat a master device moves from the device implementation to the
parent class.
Currently, device types like Bond hack around ignore-carrier
setting, as they always want to ignore-carrier.
Prepare so that also for such master types, we rely and honor the
ignore-carrier setting better. In the next commit, bond, bridge and
team devices they will get ignore-carrier turned on by default.
When a user forces up a connection on a device, mark earlier the
device as managed: this would allow proper clean-up on the device also
when it was previously unmanaged or assumed.
This would avoid skipping IPv6LL address generation when instead it was
needed.
Fixes: adbf383628https://bugzilla.redhat.com/show_bug.cgi?id=1452046
These properties are internal and shall not be publicly accessible.
Remove the getter.
We may later no longer use GDBusObjectManager. It should be an implementation
detail, not exposed in the public API of NMObject.
The current implementation of GDBusObjectManagerClient implements
GAsyncInitableIface, however it simply runs GInitableIface's
synchronous init on another thread.
I suspect that there are races in the way that is implemented.
For one, we see crashes and warnings (rh#1450075, rh#1457769,
rh#1457223). Also, it seems very wrong to me, how GDBusObjectManagerClient
mixes asynchronous signals (on_control_proxy_g_signal) with
synchronously getting all objects (process_get_all_result,
GetManagedObjects).
I think we should ditch GDBusObjectManager altogether, including the
gdbus-codegen skeletons. They add layers of code, for something that
should be simple to do directly. For now, just don't do asynchronous
initialization on another thread, so we at least avoid this kind of
multithreadding issue.
This may make the initialization of NMClient a bit slower.
In _internal_activate_device(), we try to find an existing master AC
for the slave AC, and we create a new one in case of failure. The
master AC may already exist, but it may not be detected by
find_master() because it is undergoing authorization.
The result is that we auto-activate the master when there is already a
user activation in place, and the auto-activation will cancel the user
one. This is bad, as user-activation should always have precedence.
To fix this, introduce a last-minute check before activating internal
connections.
https://bugzilla.redhat.com/show_bug.cgi?id=1450219
No need to clone the list anymore. Unfortunately, GPtrArray is not NULL
terminated (without extra effort), so we have to pass on the GPtrArray
instance for the length.
