Reapply() is supposed to make sure that the system (the interface)
is configured as indicated by the applied-connection. That means,
it will remove/add configuration to make the system match the requested
configuration.
Add a flag "preserve-external-ip" which relaxes this. During reapply,
IP addresses/routes that exist on the interface and which are not known
(or added) by NetworkManager will be left alone.
This will be used by nm-cloud-setup, so that it can reconfigure the
interface in a less destructive way, which does not conflict with
external `ip addr/route` calls.
Note that the previous commit just adds "VersionInfo" and the
possibility to expose capabilities (patch-level). This is not used
for the new reapply flag, because, while we might backport the
reapply flag, we won't backport the "VersionInfo" property. Exposing
new capabilities via the "VersionInfo" property will only become useful
in the future, where we can backport a capability to older NM versions
(but those that have "VersionInfo" too).
This exposes NM_VERSION as number (contrary to the "Version", which is a
string). That is in particular useful, because the number can be
compared with <> due to the encoding of the version.
While at it, don't make it a single number. Expose an array of numbers,
where the following numbers are a bitfield of capabilities.
Note that before commit 3c67a1ec5e ('cli: remove version check against
NM'), we used to parse the "Version" string to detect the version. As
such, the information that "VersionInfo" exposes now, was already
(somewhat) available, you just had to parse the string. The main benefit of
"VersionInfo" is that it can expose capabilities (patched behavior) in
in a lightweight bitfield. To include the numerical version there is
just useful on top.
Currently no additional capabilities are exposed. The idea is of course
to have a place in the future, where we can expose additional
capabilities. Adding a capability flag is most useful for behavior that we
backport to older branches. Otherwise, we could just check the daemon version
alone. But since we only add "VersionInfo" property only now, we cannot backport
any capability further than this, because the "VersionInfo" property itself
won't be backported. As such, this will only be useful in the future by having
a place where we can add (and backport) capabilities.
Note that there is some overlap with the existing "Capability" property
and NMCapability enum. The difference is that adding a capability via "VersionInfo"
is only one bit, and thus cheaper. Most importantly, having it cheaper means
the downsides of adding a capability flag is significantly removed. In
practice, we could live without capabilities for a long time, so they
must be very cheap for them to be worth to add. Another difference might be,
that we will want that the VersionInfo is about compile time defaults (e.g.
a certain patch/behavior that is in or not), while NM_CAPABILITY_TEAM depends on
whether the team plugin is loaded at runtime.
Introduce a "vlan.protocol" property that specifies the protocol of a
VLAN, which controls the tag (EtherType) used for encapsulation.
Regular VLANs use 802.1Q (tag 0x8100). To implement VLAN stacking it's
sometimes useful to have 802.1ad VLANs with tag 0x88A8.
The property is a string instead of e.g. an enum because this allows
maximum flexibility in the future. For example, it becomes possible to
specify an arbitrary number in case if the kernel ever allows it.
Add a new "ovs-port.trunks" property that indicates which VLANs are
trunked by the port.
At ovsdb level the property is just an array of integers; on the
command line, ovs-vsctl accepts ranges and expands them.
In NetworkManager the ovs-port setting stores the trunks directly as a
list of ranges.
The next commit is going to introduce a new object in libnm to
represent a range of ovs-port VLANs. A "range of integers" object
seems something that can be used for other purposes in the future, so
instead of adding an object specific for this case
(e.g. NMOvsPortVlanRange), introduce a generic NMRange object that
generically represents a range of non-negative integers.
Support managing the loopback interface through NM as the users want to
set the proper mtu for loopback interface when forwarding the packets.
Additionally, the IP addresses, DNS, route and routing rules are also
allowed to configure for the loopback connection profiles.
https://bugzilla.redhat.com/show_bug.cgi?id=2060905
See wpa_supplicant commit [1]:
macsec: Make pre-shared CKN variable length
IEEE Std 802.1X-2010, 9.3.1 defines following restrictions for
CKN:
"MKA places no restriction on the format of the CKN, save that it
comprise an integral number of octets, between 1 and 32
(inclusive), and that all potential members of the CA use the same
CKN. No further constraints are placed on the CKNs used with PSKs,
..."
Hence do not require a 32 octet long CKN but instead allow a
shorter CKN to be configured.
This fixes interoperability with some Aruba switches, that do not
accept a 32 octet long CKN (only support shorter ones).
[1] https://w1.fi/cgit/hostap/commit/?id=b678ed1efc50e8da4638d962f8eac13312a4048f
Most users included this by accident, by including nm-connection.h. That
is not too great, becuase stuff it contains is by no means specific to
NMConnection.
Anyways, it's not like it would matter too that. I mainly care about it
being included in NetworkManager.h, so that there's one less special
case in a test that makes sure useful stuff from NetworkManager.h ends up
in gtk-doc (a separate commit).
Makes gtk-doc grumpy (but it likes getting grumpy too often for us to
actually pay attention, it seems):
libnm-core-impl/nm-utils.c:4342: warning: nm_utils_is_uuid is deprecated
in the inline comments, but no deprecation guards were found around
the declaration. (See the --deprecated-guards option for gtkdoc-scan.)
libnm-client-impl/nm-device-ovs-bridge.c:36: warning:
nm_device_ovs_bridge_get_slaves is deprecated in the inline comments,
...
libnm-client-impl/nm-device-ovs-port.c:36: warning:
nm_device_ovs_port_get_slaves is deprecated in the inline comments,
...
libnm-client-impl/nm-device-team.c:77: warning:
nm_device_team_get_slaves is deprecated in the inline comments,
...
"gen-metadata-nm-settings-libnm-core.xml" now contains also the names of
the NMSetting types, like "NMSettingConnection". That can be useful
to create NMSetting instances generically (that is, without knowing
the C API that gets called).
So you might be tempted to run
#!/bin/python
import gi
gi.require_version("NM", "1.0")
from gi.repository import GObject, NM
connection = NM.SimpleConnection()
# NM.utils_ensure_gtypes()
gtype_name = "NMSetting6Lowpan"
gtype = GObject.type_from_name(gtype_name)
setting = GObject.new(gtype)
connection.add_setting(setting)
However, without NM.utils_ensure_gtypes() that would not work, because
the GType is not yet created. For a user who doesn't know a priory all
setting types, it's not entirely clear how to make this work. Well, a
GObject introspection user could iterate over al NM.Setting* names and
try to instantiate the classes. However, that is still cumbersome, and not
accessible to a C user (without GI) and the currently loaded libnm
library may be newer and have unknown setting types.
In particular plain C user would need to know to call all the right
nm_setting_*_get_type(), functions, so it needs to know all the existing
52 type getters (and cannot support those from a newer libnm version).
With nm_utils_ensure_gtypes(), the user can get the typename and create
instances generically only using g_type_from_name().
Possible alternatives:
- libnm also has _nm_utils_init() which runs as __attribute__((constructor)).
We could also always instantiate all GType there. However, I don't like running
non-trivial, absolutely necessary code before main().
- hook nm_setting_get_type() to create all GType for the NMSetting
subclasses too. The problem is, that it's not entirely trivial to
avoid deadlock.
- hook nm_connection_get_type() to create all NMSetting types. That
would not deadlock, but it still is questionable whether we should
automatically, at non-obvious times instantiate all GTypes.
These are present in a public header yet are not properly commented,
versioned or exported.
Export them now. Another option would be to move them to a private
header; but I suspect someone has intended them to be exported at some
point.
Add them to @libnm_1_40_4 as opposed to @libnm_1_42_0 because we now know
this is going to be backported to 1.40.4 first.
In the commit 2a11c57c4e ('libnm/wifi: rework NMSetting8021xAuthFlags
to explicitly disable TLS version'), it said:
> In the future, supplicant may disable options by default, and
> the inverse option can become interesting to configure
> "tls_disable_tlsv1_0=0". When that happens, we can solve it by
> adding another flag NM_SETTING_802_1X_AUTH_FLAGS_TLS_1_0_ENABLE.
This commit adds the `NM_SETTING_802_1X_AUTH_FLAGS_TLS_1_0_ENABLE`
flag as well as similar flags for other TLS versions.
This commit also adds flags for TLS v1.3, as the corresponding flags
are now provided in wpa_supplicant.
The NMSetting8021xAuthFlags setting is rejected when both enable and
disable are set for the same TLS version. if-else-if is used in
nm_supplicant_config_add_setting_8021x to guarantee this behavior.
It prefers ENABLE over DISABLE to match the behavior of wpa_supplicant.
https://gitlab.freedesktop.org/NetworkManager/NetworkManager/-/issues/1133https://gitlab.freedesktop.org/NetworkManager/NetworkManager/-/merge_requests/1450
- the static assertions were wrong, there was a "," instead of "==".
- the numeric values were wrong, as shown by the static assertions.
- move the code comment to the implementation. This does not seem
relevant for the library user and should not be in the public header.
Fixes: 08e845f651 ('nm-setting: mangle public constant to make g-ir-scanner happy')
This is part of public API, and was wrongly renamed during some internal
refactoring.
Reported-by: Eivind Næss <eivnaes@yahoo.com>
Fixes: 08eff4c46e ('glib-aux: rename IP address related helpers from "nm-inet-utils.h"')
Add option to set ofport_request when configuring ovs interface. When
connection with ofport_request configured is activated ovsdb will first
try to activated on the port set by ofport_request.
the .h.in file is not formatted by our nm-code-format.sh
file. It also contains .in template parameters that the
formatting would destroy.
Still, follow our current style and reformat the parts manually.
1) The "enabled-on-global-iface" flag was odd. Instead, have only
and "enabled" flag and skip (by default) endpoints on interface
that have no default route. With the new flag "also-without-default-route",
this can be overruled. So previous "enabled-on-global-default" now is
the same as "enabled", and "enabled" from before behaves now like
"enabled,also-without-default-route".
2) What was also odd, as that the fallback default value for the flags
depends on "/proc/sys/net/mptcp/enabled". There was not one fixed
fallback default, instead the used fallback value was either
"enabled-on-global-iface,subflow" or "disabled".
Usually that is not a problem (e.g. the default value for
"ipv6.ip6-privacy" also depends on use_tempaddr sysctl). In this case
it is a problem, because the mptcp-flags (for better or worse) encode
different things at the same time.
Consider that the mptcp-flags can also have their default configured in
"NetworkManager.conf", a user who wants to switch the address flags
could previously do:
[connection.mptcp]
connection.mptcp-flags=0x32 # enabled-on-global-iface,signal,subflow
but then the global toggle "/proc/sys/net/mptcp/enabled" was no longer
honored. That means, MPTCP handling was always on, even if the sysctl was
disabled. Now, "enabled" means that it's only enabled if the sysctl
is enabled too. Now the user could write to "NetworkManager.conf"
[connection.mptcp]
connection.mptcp-flags=0x32 # enabled,signal,subflow
and MPTCP handling would still be disabled unless the sysctl
is enabled.
There is now also a new flag "also-without-sysctl", so if you want
to really enable MPTCP handling regardless of the sysctl, you can.
The point of that might be, that we still can configure endpoints,
even if kernel won't do anything with them. Then you could just flip
the sysctl, and it would start working (as NetworkManager configured
the endpoints already).
Fixes: eb083eece5 ('all: add NMMptcpFlags and connection.mptcp-flags property')
- name things related to `in_addr_t`, `struct in6_addr`, `NMIPAddr` as
`nm_ip4_addr_*()`, `nm_ip6_addr_*()`, `nm_ip_addr_*()`, respectively.
- we have a wrapper `nm_inet_ntop()` for `inet_ntop()`. This name
of our wrapper is chosen to be familiar with the libc underlying
function. With this, also name functions that are about string
representations of addresses `nm_inet_*()`, `nm_inet4_*()`,
`nm_inet6_*()`. For example, `nm_inet_parse_str()`,
`nm_inet_is_normalized()`.
<<<<
R() {
git grep -l "$1" | xargs sed -i "s/\<$1\>/$2/g"
}
R NM_CMP_DIRECT_IN4ADDR_SAME_PREFIX NM_CMP_DIRECT_IP4_ADDR_SAME_PREFIX
R NM_CMP_DIRECT_IN6ADDR_SAME_PREFIX NM_CMP_DIRECT_IP6_ADDR_SAME_PREFIX
R NM_UTILS_INET_ADDRSTRLEN NM_INET_ADDRSTRLEN
R _nm_utils_inet4_ntop nm_inet4_ntop
R _nm_utils_inet6_ntop nm_inet6_ntop
R _nm_utils_ip4_get_default_prefix nm_ip4_addr_get_default_prefix
R _nm_utils_ip4_get_default_prefix0 nm_ip4_addr_get_default_prefix0
R _nm_utils_ip4_netmask_to_prefix nm_ip4_addr_netmask_to_prefix
R _nm_utils_ip4_prefix_to_netmask nm_ip4_addr_netmask_from_prefix
R nm_utils_inet4_ntop_dup nm_inet4_ntop_dup
R nm_utils_inet6_ntop_dup nm_inet6_ntop_dup
R nm_utils_inet_ntop nm_inet_ntop
R nm_utils_inet_ntop_dup nm_inet_ntop_dup
R nm_utils_ip4_address_clear_host_address nm_ip4_addr_clear_host_address
R nm_utils_ip4_address_is_link_local nm_ip4_addr_is_link_local
R nm_utils_ip4_address_is_loopback nm_ip4_addr_is_loopback
R nm_utils_ip4_address_is_zeronet nm_ip4_addr_is_zeronet
R nm_utils_ip4_address_same_prefix nm_ip4_addr_same_prefix
R nm_utils_ip4_address_same_prefix_cmp nm_ip4_addr_same_prefix_cmp
R nm_utils_ip6_address_clear_host_address nm_ip6_addr_clear_host_address
R nm_utils_ip6_address_same_prefix nm_ip6_addr_same_prefix
R nm_utils_ip6_address_same_prefix_cmp nm_ip6_addr_same_prefix_cmp
R nm_utils_ip6_is_ula nm_ip6_addr_is_ula
R nm_utils_ip_address_same_prefix nm_ip_addr_same_prefix
R nm_utils_ip_address_same_prefix_cmp nm_ip_addr_same_prefix_cmp
R nm_utils_ip_is_site_local nm_ip_addr_is_site_local
R nm_utils_ipaddr_is_normalized nm_inet_is_normalized
R nm_utils_ipaddr_is_valid nm_inet_is_valid
R nm_utils_ipx_address_clear_host_address nm_ip_addr_clear_host_address
R nm_utils_parse_inaddr nm_inet_parse_str
R nm_utils_parse_inaddr_bin nm_inet_parse_bin
R nm_utils_parse_inaddr_bin_full nm_inet_parse_bin_full
R nm_utils_parse_inaddr_prefix nm_inet_parse_with_prefix_str
R nm_utils_parse_inaddr_prefix_bin nm_inet_parse_with_prefix_bin
R test_nm_utils_ip6_address_same_prefix test_nm_ip_addr_same_prefix
./contrib/scripts/nm-code-format.sh -F
It can be useful to choose a different "ipv6.addr-gen-mode". And it can be
useful to override the default for a set of profiles.
For example, in cloud or in a data center, stable-privacy might not be
the best choice. Add a mechanism to override the default via global defaults
in NetworkManager.conf:
# /etc/NetworkManager/conf.d/90-ipv6-addr-gen-mode-override.conf
[connection-90-ipv6-addr-gen-mode-override]
match-device=type:ethernet
ipv6.addr-gen-mode=0
"ipv6.addr-gen-mode" is a special property, because its default depends on
the component that configures the profile.
- when read from disk (keyfile and ifcfg-rh), a missing addr-gen-mode
key means to default to "eui64".
- when configured via D-Bus, a missing addr-gen-mode property means to
default to "stable-privacy".
- libnm's ip6-config::addr-gen-mode property defaults to
"stable-privacy".
- when some tool creates a profile, they either can explicitly
set the mode, or they get the default of the underlying mechanisms
above.
- nm-initrd-generator explicitly sets "eui64" for profiles it creates.
- nmcli doesn' explicitly set it, but inherits the default form
libnm's ip6-config::addr-gen-mode.
- when NM creates a auto-default-connection for ethernet ("Wired connection 1"),
it inherits the default from libnm's ip6-config::addr-gen-mode.
Global connection defaults only take effect when the per-profile
value is set to a special default/unset value. To account for the
different cases above, we add two such special values: "default" and
"default-or-eui64". That's something we didn't do before, but it seams
useful and easy to understand.
Also, this neatly expresses the current behaviors we already have. E.g.
if you don't specify the "addr-gen-mode" in a keyfile, "default-or-eui64"
is a pretty clear thing.
Note that usually we cannot change default values, in particular not for
libnm's properties. That is because we don't serialize the default
values to D-Bus/keyfile, so if we change the default, we change
behavior. Here we change from "stable-privacy" to "default" and
from "eui64" to "default-or-eui64". That means, the user only experiences
a change in behavior, if they have a ".conf" file that overrides the default.
https://bugzilla.redhat.com/show_bug.cgi?id=1743161https://bugzilla.redhat.com/show_bug.cgi?id=2082682
See-also: https://github.com/coreos/fedora-coreos-tracker/issues/907https://gitlab.freedesktop.org/NetworkManager/NetworkManager/-/merge_requests/1213
The property wait-activation-delay will delay the activation of an
interface the specified amount of milliseconds. Please notice that it
could be delayed some milliseconds more due to other events in
NetworkManager.
This could be used in multiple scenarios where the user needs to define
an arbitrary delay e.g LACP bond configure where the LACP negotiation
takes a few seconds and traffic is not allowed, so they would like to
use nm-online and a setting configured with this new property to wait
some seconds. Therefore, when nm-online is finished, LACP bond should be
ready to receive traffic.
The delay will happen right before the device is ready to be activated.
https://gitlab.freedesktop.org/NetworkManager/NetworkManager/-/merge_requests/1248https://bugzilla.redhat.com/show_bug.cgi?id=2008337
Introduction of a new setting ipv4.link-local, which enables
link-local IP addresses concurrently with other IP address assignment
implementations such as dhcp or manually.
No way is implemented to obtain a link-local address as a fallback when
dhcp does not respond (as dhcpd does, for example). This could be be
added later.
To maintain backward compatibility with ipv4.method ipv4.link-local has
lower priority than ipv4.method. This results in:
* method=link-local overrules link-local=disabled
* method=disabled overrules link-local=enabled
Furthermore, link-local=auto means that method defines whether
link-local is enabled or disabled:
* method=link-local --> link-local=enabled
* else --> link-local=disabled
The upside is, that this implementation requires no normalization.
Normalization is confusing to implement, because to get it really
right, we probably should support normalizing link-local based on
method, but also vice versa. And since the method affects how other
properties validate/normalize, it's hard to normalize that one, so that
the result makes sense. Normalization is also often not great to the
user, because it basically means to modify the profile based on other
settings.
The downside is that the auto flag becomes API and exists because
we need backward compatibility with ipv4.method.
We would never add this flag, if we would redesign "ipv4.method"
(by replacing by per-method-specific settings).
Defining a default setting for ipv4.link-local in the global
configuration is also supported.
The default setting for the new property can be "default", since old
users upgrading to a new version that supports ipv4.link-local will not
have configured the global default in NetworkManager.conf. Therefore,
they will always use the expected "auto" default unless they change
their configuration.
Co-Authored-By: Thomas Haller <thaller@redhat.com>
libnm-core is also used by the daemon, thus currently dragging in
libnm-crypto there. But could we ever drop that dependency?
One use of the libnm-crypto is in functions like nm_utils_file_is_certificate()
in "nm-utils.h". These are part of the public API of libnm.
But this is not used by the daemon. Move it to "libnm-client-core"
to be closer to where it's actually used.
As we have unit tests in "libnm-core-impl/tests" that test this function,
those unit tests also would need to move to "libnm-client-impl".
Instead, add the actual implementation of these function to "libnm-crypto"
and test it there.
This patch moves forward declarations from public header "nm-utils.h" to
"nm-client.h". Arguably, "nm-client.h" is not a great name, but we don't
have a general purpose header in "libnm-client-public", so use this.
Note that libnm users can only include <NetworkManager.h> and including
individual files is not supported (and even prevented). Thus moving
the declarations won't break any users.
Introduce a RadioFlags property on the manager object. For now it
contains two bits WLAN_AVAILABLE, WWAN_AVAILABLE to indicate whether
any radio interface is present in the system. The presence of a radio
is detected by looking at devices and rfkill switches.
In future, any radio-related read-only boolean flag can be exposed via
this property, including the already existing WirelessHardwareEnabled
and WwanHardwareEnabled properties.
When we have a bridge interface with ports attached externally (that is,
not by NetworkManager itself), then it can make sense that during
checkpoint rollback we want to keep those ports attached.
During rollback, we may need to deactivate the bridge device and
re-activate it. Implement this, by setting a flag before deactivating,
which prevents external ports to be detached. The flag gets cleared,
when the device state changes to activated (the following activation)
or unmanaged.
This is an ugly solution, for several reasons.
For one, NMDevice tracks its ports in the "slaves" list. But what
it does is ugly. There is no clear concept to understand what it
actually tacks. For example, it tracks externally added interfaces
(nm_device_sys_iface_state_is_external()) that are attached while
not being connected. But it also tracks interfaces that we want to attach
during activation (but which are not yet actually enslaved). It also tracks
slaves that have no actual netdev device (OVS). So it's not clear what this
list contains and what it should contain at any point in time. When we skip
the change of the slaves states during nm_device_master_release_slaves_all(),
it's not really clear what the effects are. It's ugly, but probably correct
enough. What would be better, if we had a clear purpose of what the
lists (or several lists) mean. E.g. a list of all ports that are
currently, physically attached vs. a list of ports we want to attach vs.
a list of OVS slaves that have no actual netdev device.
Another problem is that we attach state on the device
("activation_state_preserve_external_ports"), which should linger there
during the deactivation and reactivation. How can we be sure that we don't
leave that flag dangling there, and that the desired following activation
is the one we cared about? If the follow-up activation fails short (e.g. an
unmanaged command comes first), will we properly disconnect the slaves?
Should we even? In practice, it might be correct enough.
Also, we only implement this for bridges. I think this is where it makes
the most sense. And after all, it's an odd thing to preserve unknown,
external things during a rollback -- unknown, because we have no knowledge
about why these ports are attached and what to do with them.
Also, the change doesn't remember the ports that were attached when the
checkpoint was created. Instead, we preserve all ports that are attached
during rollback. That seems more useful and easier to implement. So we
don't actually rollback to the configuration when the checkpoint was
created. Instead, we rollback, but keep external devices.
Also, we do this now by default and introduce a flag to get the previous
behavior.
https://bugzilla.redhat.com/show_bug.cgi?id=2035519https://gitlab.freedesktop.org/NetworkManager/NetworkManager/-/issues/ # 909
The problem was that glib-mkenums requires all enum values on a separate
line. And clang-format would put all on the same line, unless the last
value has a trailing comma. Which is the better solution here.
