Instead of throwing an assertion, fail DHCPv6 when a IPv6 link-local
address is not configured on the device. There are different reasons
why the assertion may fail: for example the address was removed
externally; or the device is gone (and thus the platform already
received the notification of addresses removal) but the device is still
connecting because its disposal happens in an idle callback.
None of these deserves an assertion, which should only be for
programming errors.
https://bugzilla.redhat.com/show_bug.cgi?id=1432251
Udev may be slow at startup and it may take a while until the
device is visible in udev. Before that, there are no pending
actions yet because the device is still in unmanaged state.
Hack nm_device_has_pending_action() to indicate a pending action
when the platform link is not yet initialized.
We don't bother using nm_device_add_pending_action() to schedule
a proper pending-action. It is simpler this way, also we precisely
log about the state of NM_UNMANAGED_PLATFORM_INIT flag. The pending
actions are implemented in their way mostly for logging purpose to
understand what blocks a device. For NM_UNMANAGED_PLATFORM_INIT we
have sufficient logging so no need for the overhead and effort.
https://bugzilla.gnome.org/show_bug.cgi?id=779920
In order to support a reapply of proxy configuration, remember the PAC
URL received through DHCP and merge it again with configuration from
proxy setting on reapply.
Add two new methods can_reapply_change() and reapply_connection() to
the device class. In this way device subclasses can declare that they
support the changes in a given setting and take care of reapplying it.
The state-change of a device has a reason argument, which is mostly for information
only.
There are many places in code that are the source of a state-reason.
Mostly these are calls to:
- nm_device_state_changed()
- nm_device_queue_state()
- nm_device_queue_recheck_available()
- nm_device_set_unmanaged_by_*()
- nm_device_master_release_one_slave()
- nm_device_ip_method_failed()
- nm_modem_emit_prepare_result()
- nm_modem_emit_ppp_failed()
- nm_manager_deactivate_connection()
- NM_SET_OUT (out_failure_reason, NM_DEVICE_STATE_REASON_*);
However, there are a few places in code that look at the reason
to decide how to proceed. I think this is a bad pattern, because
cause and effect are decoupled and it gets hard to understand where
a certain reason is set and what consequences that has.
Add a nop-function nm_device_state_reason_check() to mark all uses
of the device state reason that derive decisions from it. That is,
highlight the "effect" part.
The out-reason is only set to NM_DEVICE_STATE_REASON_CONFIG_FAILED.
And there is only one caller who cares about the reason.
If we one day decide to return a more distinguished error reasons,
we can revert this patch. Until then, drop the code.
This argument is only relevant when the NMActStageReturn argument
indicates NM_ACT_STAGE_RETURN_FAILURE. In all other cases it is ignored.
Rename the argument to make the meaning clearer. The argument is passed
through several layers of code, it isn't obvious that this argument only
matters for the failure case. Also, the distinct name makes it easier
to distinguish from other uses of the "reason" name.
While at it, do some drive-by cleanup:
- use g_return_*() instead of g_assert() to have a more graceful
assertion.
- functions like dhcp4_start() don't need to return a failure reason.
Most callers don't care, and the caller who does can determine the
proper reason.
- allow omitting the out-argument via NM_SET_OUT().
Add support for creating dummy devices. This commit adds a D-Bus
interface 'org.freedesktop.NetworkManager.Device.Dummy' which is used
primarily for determining the device type but does not carry any
properties.
Use the per-connection authentication timeout for 802.1X Ethernet,
MACsec and Wi-Fi connections. In case the value is not defined, fall
back to the global one.
It allows derived classes to override the autoconnect-allowed
state.
We already have
- NM_DEVICE_AUTOCONNECT property, which is two parts:
- NMDevicePrivate::autoconnect_user, which is settable via
D-Bus by the use, to allow the device to autoconnect.
- NMDevicePrivate::autoconnect_intern, which is set by
internal decision.
- NM_DEVICE_AUTOCONNECT_ALLOWED signal, where other devices can
subscribe to block autoconnect. Currently that is only used
by NMDeviceOlpcMesh.
These two make up for nm_device_autoconnect_allowed().
Add another way to allow derived classes to disable autoconnect
temporarily. This could also be achieved by having the device
subscribe to NM_DEVICE_AUTOCONNECT_ALLOWED of self, or by adding
a signal slot. But a plain function pointer seems easier.
The NMDevice's autoconnect property is settable via D-Bus and is is
also modified by internal decision, like when no PIN is available.
Certain internal actions cause clearing the internal autoconnect flag,
but they should not override the user desicion.
For example, when NM awaks from sleep it would reenable autoconnect,
but it should not reenable it for devices where the user explicitly
said that autoconnect is to be disabled.
Similarly, activating a device alone is not yet an instruction to
re-enable autoconnect. If the user consciously disables autoconnect,
it should stay enabled. On the other hand, activating a device should
reenable autoconnect if it was blocked by internal decision.
We need to track these two flags separately, and set them accordingly.
- no longer bother clearing .state and .reason when the .id
field is unset. The fields just don't matter and no user
accesses these fields when the glib source id is not set.
- unify logging and give them all a prefix "queue-state[%s, %s, %u]: ".
- drop nm_device_queued_state_peek(), it only had one caller,
thus inline the trivial check.
- make nm_device_queued_state_clear() a static function
queued_state_clear()
- rename queued_set_state() to queued_state_set().
Reorder code to be like in other source files:
- first includes and generic defines
- then various helper structs
- then GObject related declarations, with first signal and property
enums, then the private data, then the G_DEFINE_TYPE() itself.
- finally, forward declarations for functions.
We don't want to waste a full "int" size to store the @hw_addr_type
in NMDevicePrivate. Previously, that was hacked around by using guint8.
Now, instead use a bitfield which has the right type.
These two structs are only used at exactly one place: as the type
for a field in NMDevicePrivate.
Having additional structs (that are only used at one place) only
add noise. Also, there are already prior-acts of using unnamed
structs in NMDevicePrivate in case of structs that only serve
to group/namespace a set of fields.
All callers either use a static @action argument or keep a clone
of the string that lives as long as the action is pending. So,
save cloning the string.
Startup-complete means that all devices have settled in a state
and no further activation is pending. When we have a recheck-available
scheduled, we clearly should not yet declare startup-complete.
Add a new pending-action "recheck-available" to avoid:
<info> [1485520408.3920] device (wlp2s0): supplicant interface state: starting -> ready
<debug> [1485520408.3920] device[0x563abbcca400] (wlp2s0): remove_pending_action (0): 'waiting for supplicant'
<info> [1485520408.3920] manager: startup complete
<debug> [1485520408.3924] device[0x563abbcca400] (wlp2s0): add_pending_action (1): 'queued state change to disconnected'
The -Wimplicit-fallthrough=3 warning is quite flexible of accepting
a fall-through warning.
Some comments were missing or not detected correctly.
Thereby, also change all other comments to follow the exact
same pattern.
The public property NM_DEVICE_ACTIVATION_REQUEST exposes the exported
D-Bus path. So, it's not sufficient to emit property changed signals
when changing the priv->act_request pointer, we must also react on
exporting/unexporting.
It's not clear whether this fixes an actual bug. Maybe, we never
export/unexport priv->act_request while the device tracks it.
But the code is pretty hard to follow and it's hard to verify
whether this is the case.
By hooking up to "notify::path", we can easily verify that such
a situtation cannot arise.
(cherry picked from commit 9ae5e6a54d)
Previously, we would have different functions like
- nm_match_spec_device_type()
- nm_match_spec_hwaddr()
- nm_match_spec_s390_subchannels()
- nm_match_spec_interface_name()
which all would handle one type of match-spec.
So, to get the overall result whether the arguments
match or not, nm_device_spec_match_list() had to stich
them together and iterate the list multiple times.
Refactor the code to have one nm_match_spec_device()
function that gets all relevant paramters.
The upside is:
- the logic how to evaluate the match-spec is all at one place
(match_device_eval()) instead of spread over multiple
functions.
- It requires iterating the list at most twice. Twice, because
we do a fast pre-search for "*".
One downside could be, that we have to pass all 4 arguments
for the evaluation, even if the might no be needed. That is,
because "nm-core-utils.c" shall be independend from NMDevice, it
cannot receive a device instance to get the parameters as needed.
As we would add new match-types, the argument list would grow.
However, all arguments are cached and fetching them from the
device's private data is very cheap.
(cherry picked from commit b957403efd)
After commit 553717bb1c ("device: don't set ip4_state=IP_FAIL for
ipv4.method=disabled"), we commit an empty IPv4 configuration when
IPv4 is disabled. This means that it's not necessary anymore to call
_commit_mtu() because the MTU will be set in
ip4_config_merge_and_apply().
(cherry picked from commit 714b18dcf7)
Basically to silence gcc that is not smart enough to understand how does
.initialized and .value relate.
src/devices/nm-device.c: In function '_commit_mtu':
src/devices/nm-device.c:6754:15: error: 'ip6_mtu_sysctl.value' may be used uninitialized in this function [-Werror=maybe-uninitialized]
if (ip6_mtu && ip6_mtu != _IP6_MTU_SYS ()) {
^
(cherry picked from commit 7ce805d49d)
This allows a user to restore the previous behavior where NetworkManager
would not reconfigure the MTU during device activation, if no MTU is
available (commit "22e8af6 device: set a per-device default MTU on
activation").
Well, not exactly. The previous behavior was to use per-connection
configuration, then DHCP provided value, or finally leave the MTU
unspecified.
Now, we prefer a per-connection configuration, followed by a global
connection default. If "ethernet.mtu=0", the MTU is left unspecified.
In absense of a global connection default, the value from DHCP is used
or finally a per-device-type default. That is effectively 1500 for most
types, except for infiniband where the MTU is still left unspecified.
In absence of an explicit MTU (either via user configuration, PPP or
DHCP), set a default MTU on activation that depends on the device type.
We only want to do that on the very first call to _commit_mtu(). Later
calls (for example in response to new DHCP leases) skip over this step.
This means, on activation the MTU will always be reset to a sensible
value instead of preserving whatever was left from a previous
configuration.
This does not cover setting the MTU from the VPN plugin :(
When you have a connection with "ethernet.mtu=0 (auto)", the MTU is not set
during activation. That means, the effective MTU depends on the previous
MTU configuration of the device. Which in turn, depends on the
previously active connection, as we don't reset the MTU on deactivation.
Restore the previous MTU on deactivation iff NetworkManager changed
the MTU during device activation.
Don't have this mtu_desired variable. All the data is readily available
without redundancy. E.g. the applied-connection contains everything
we need to know. Just get it as needed.
Also drop apply_mtu_from_config(). It didn't take into account
the MTU settings beside NMSettingWired.
Also, no longer merge the NM_IP_CONFIG_SOURCE_USER MTU value into
priv->ip4_config. NMIP4Config now only tracks the MTU from the various
non-user-config sources, but the user config is no longer merged back
into the composite.
Instead of overwriting ip4_config_pre_commit(), add a new function
get_mtu().
This also adds a default value in case there is no user-configuration.
This will allow us later to reset a default MTU based on the device
type.
