As of upstream kernel v4.18-rc8.
Note that we name the features like they are called in ethtool's
ioctl API ETH_SS_FEATURES.
Except, for features like "tx-gro", which ethtool utility aliases
as "gro". So, for those features where ethtool has a built-in,
alternative name, we prefer the alias.
And again, note that a few aliases of ethtool utility ("sg", "tso", "tx")
actually affect more than one underlying kernel feature.
Note that 3 kernel features which are announced via ETH_SS_FEATURES are
explicitly exluded because kernel marks them as "never_changed":
#define NETIF_F_NEVER_CHANGE (NETIF_F_VLAN_CHALLENGED | \
NETIF_F_LLTX | NETIF_F_NETNS_LOCAL)
We will add a large number of offload features. That means, the output
of `nmcli connection show "$PROFILE"` would be very verbose, in case
the profile has a [ethtool] option.
Since this is newly added API, don't do that. Don't show ethtool properties
that are left unset.
A minor problem here is, that it becomes no longer obvious which
properties exist. We should however counter that by documentation.
Also, one could do:
$ nmcli connection modify "$PROFILE" ethtool.xxx x
Error: invalid property 'xxx': 'xxx' not among [feature-gro, feature-gso, feature-lro, feature-ntuple, feature-rx, feature-rxhash, feature-rxvlan, feature-sg, feature-tso, feature-tx, feature-txvlan, feature-tx-tcp6-segmentation, feature-tx-tcp-segmentation].
Likewise, bash completion still works as one would expect.
$ nmcli --complete-args connection modify "$PROFILE" ethtool.
ethtool.feature-gro
ethtool.feature-gso
ethtool.feature-lro
[...]
Note the output of
$ nmcli -f ethtool.feature-gro connection show "$PROFILE"
gives now nothing (if there is an ethtool section, but not this
particular feature). Maybe this shouldn't be like that. On the other
hand, specifying a connection setting that doesn't exist also gives
no output:
$ nmcli -f bond connection show "$PROFILE"
So, maybe this behavior is fine.
Historically, nmcli printed all fields during operations like
`nmcli connection show "$PROFILE"`. As we supported more and
more options, this resulted in a verbose output, of most properties
just being the default values.
To counter that, we added the '-overview' option. When given,
it would hide options that are set at their default. This option
was not the default, to preserve established behavior.
However, for new options, we can afford to hide them. Add a mechanism,
that property getters can mark their value to be hidden. At the moment,
there is no way to show these properties. However, we could add a
'-verbose' option, with the opposite meaning of '-overview'. Anyway,
that does not seem necessary at the moment.
Hiding properties from output is only acceptable for new properties
(otherwise we badly change behavior), and if the properties are set
at their default values (otherwise, we hide important information).
Also, add two more features "tx-tcp-segmentation" and
"tx-tcp6-segmentation". There are two reasons for that:
- systemd-networkd supports setting these two features,
so lets support them too (apparently they are important
enough for networkd).
- these two features are already implicitly covered by "tso".
Like for the "ethtool" program, "tso" is an alias for several
actual features. By adding two features that are already
also covered by an alias (which sets multiple kernel names
at once), we showcase how aliases for the same feature can
coexist. In particular, note how setting
"tso on tx-tcp6-segmentation off" will behave as one would
expect: all 4 tso features covered by the alias are enabled,
except that particular one.
There is little difference in practice because there is only one caller.
Still re-use the SocketHandle also for mii. If only, to make it clear
that SocketHandle is not only suitable for ethtool, but also mii.
Previously, each call to ethtool_get() would resolve the ifindex and
create a new socket for the ethtool request.
This is partly done, because ethtool only supports making requests by
name. Since interfaces can be renamed, this is inherrently racy. So,
we want to fetch the latest name shortly before making the request.
Some functions like nmp_utils_ethtool_supports_vlans() require multiple
ioctls. And next, we will introduce more ethtool functions, that make an
even larger number of individual requests.
Add a simple SocketHandle struct, to create the socket once and reuse
it for multiple requests. This is still entirely internal API in
"nm-platform-utils.c".
ethtool_get_stringset() will be used later, independently.
Also, don't trust and ensure that the block of strings
returned by ETHTOOL_GSTRINGS are NUL terminated.
Note that in NetworkManager API (D-Bus, libnm, and nmcli),
the features are called "feature-xyz". The "feature-" prefix
is used, because NMSettingEthtool possibly will gain support
for options that are not only -K|--offload|--features, for
example -C|--coalesce.
The "xzy" suffix is either how ethtool utility calls the feature
("tso", "rx"). Or, if ethtool utility specifies no alias for that
feature, it's the name from kernel's ETH_SS_FEATURES ("tx-tcp6-segmentation").
If possible, we prefer ethtool utility's naming.
Also note, how the features "feature-sg", "feature-tso", and
"feature-tx" actually refer to multiple underlying kernel features
at once. This too follows what ethtool utility does.
The functionality is not yet implemented server-side.
Add a new way how NMSetting subclasses can be implemented.
Currently, most NMSetting implementations realize all their properties
via GObject properties. That has some downsides:
- the biggest one, is the large effort to add new properties.
Most of them are implemented on a one-by-one basis and they come
with additional API (like native getter functions).
It makes it cumbersome to add more properties.
- for certain properties, it's hard to encode them entirely in
a GObject property. That results in unusable API like
NM_SETTING_IP_CONFIG_ADDRESSES, NM_SETTING_BOND_OPTIONS,
NM_SETTING_USER_DATA. These complex valued properties only
exist, because we currently always need GObject properties
to even implement simple functionality. For example,
nm_setting_duplicate() is entirely implemented via
nm_setting_enumerate_values(), which can only iterate
GObject properies. There is no reason why this is necessary.
Note also how nmcli badly handles bond options and VPN
data. That is only a shortcoming of nmcli and wouldn't
need to be that way. But it happend, because we didn't
keep an open mind that settings might be more than just
accessing GObject properties.
- a major point of NMSetting is to convert to/from a GVariant
from the D-Bus API. As NMSetting needs to squeeze all values
into the static GObject structure, there is no place to
encode invalid or unknown properties. Optimally,
_nm_setting_new_from_dbus() does not loose any information
and a subsequent _nm_setting_to_dbus() can restore the original
variant. That is interesting, because we want that an older
libnm client can talk to a newer NetworkManager version. The
client needs to handle unknown properties gracefully to stay
forward compatible. However, it also should not just drop the
properties on the floor.
Note however, optimally we want that nm_setting_verify() still
can reject settings that have such unknown/invalid values. So,
it should be possible to create an NMSetting instance without
error or loosing information. But verify() should be usable to
identify such settings as invalid.
They also have a few upsides.
- libnm is heavily oriented around GObject. So, we generate
our nm-settings manual based on the gtk-doc. Note however,
how we fail to generate a useful manual for bond.options.
Also note, that there is no reason we couldn't generate
great documentation, even if the properties are not GObject
properties.
- GObject properties do give some functionality like meta-data,
data binding and notification. However, the meta-data is not
sufficient on its own. Note how keyfile and nmcli need extensive
descriptor tables on top of GObject properties, to make this
useful. Note how GObject notifications for NMSetting instances
are usually not useful, aside for data binding like nmtui does.
Also note how NMSettingBond already follows a different paradigm
than using GObject properties. Nowdays, NMSettingBond is considered
a mistake (related bug rh#1032808). Many ideas of NMSettingBond
are flawed, like exposing an inferiour API that reduces everything
to a string hash. Also, it only implemented the options hash inside
NMSettingBond. That means, if we would consider this a good style,
we would have to duplicate this approach in each new setting
implementation.
Add a new style to track data for NMSetting subclasses. It keeps
an internal hash table with all GVariant properies. Also, the
functionality is hooked into NMSetting base class, so all future
subclasses that follow this way, can benefit from this. This approach
has a few similiarties with NMSettingBond, but avoids its flaws.
With this, we also no longer need GObject properties (if we would
also implement generating useful documentation based on non-gkt-doc).
They may be added as accessors if they are useful, but there is no
need for them.
Also, handling the properties as a hash of variants invites for a
more generic approach when handling them. While we still could add
accessors that operate on a one-by-one bases, this leads to a more
generic usage where we apply common functionality to a set of properties.
Also, this is for the moment entirely internal and an implementation
detail. It's entirely up to the NMSetting subclass to make use of this
new style. Also, there are little hooks for the subclass available.
If they turn out to be necessary, they might be added. However, for
the moment, the functionality is restricted to what is useful and
necessary.
NMSetting internally already tracked a list of all proper GObject properties
and D-Bus-only properties.
Rework the tracking of the list, so that:
- instead of attaching the data to the GType of the setting via
g_type_set_qdata(), it is tracked in a static array indexed by
NMMetaSettingType. This allows to find the setting-data by simple
pointer arithmetic, instead of taking a look and iterating (like
g_type_set_qdata() does).
Note, that this is still thread safe, because the static table entry is
initialized in the class-init function with _nm_setting_class_commit().
And it only accessed by following a NMSettingClass instance, thus
the class constructor already ran (maybe not for all setting classes,
but for the particular one that we look up).
I think this makes initialization of the metadata simpler to
understand.
Previously, in a first phase each class would attach the metadata
to the GType as setting_property_overrides_quark(). Then during
nm_setting_class_ensure_properties() it would merge them and
set as setting_properties_quark(). Now, during the first phase,
we only incrementally build a properties_override GArray, which
we finally hand over during nm_setting_class_commit().
- sort the property infos by name and do binary search.
Also expose this meta data types as internal API in nm-setting-private.h.
While not accessed yet, it can prove beneficial, to have direct (internal)
access to these structures.
Also, rename NMSettingProperty to NMSettInfoProperty to use a distinct
naming scheme. We already have 40+ subclasses of NMSetting that are called
NMSetting*. Likewise, NMMetaSetting* is heavily used already. So, choose a
new, distinct name.
We have NMMetaSettingType enum, which is an enum of all setting types.
We also have an efficient way to get the enum (and its NMMetaSettingInfo)
from an NMSetting, setting-name and GType.
No longer maintain the vtable for keyfile by "setting-name". Instead,
index it by NMMetaSettingType enum.
That way, we get efficient lookup, and don't need to duplicate the
functionality of finding the vtable entry for a setting.
nm_meta_setting_infos_by_name() did a naive search by name by
iterating over all 42 setting types.
Reorder nm_meta_setting_infos array, and use binary search instead.
Previously, each (non abstract) NMSetting class had to register
its name and priority via _nm_register_setting().
Note, that libnm-core.la already links against "nm-meta-setting.c",
which also redundantly keeps track of the settings name and gtype
as well.
Re-use NMMetaSettingInfo also in libnm-core.la, to track this meta
data.
The goal is to get rid of private data structures that track
meta data about NMSetting classes. In this case, "registered_settings"
hash. Instead, we should have one place where all this meta data
is tracked. This was, it is also accessible as internal API,
which can be useful (for keyfile).
Note that NMSettingClass has some overlap with NMMetaSettingInfo.
One difference is, that NMMetaSettingInfo is const, while NMSettingClass
is only initialized during the class_init() method. Appart from that,
it's mostly a matter of taste, whether we attach meta data to
NMSettingClass, to NMMetaSettingInfo, or to a static-array indexed
by NMMetaSettingType.
Note, that previously, _nm_register_setting() was private API. That
means, no user could subclass a functioning NMSetting instance. The same
is still true: NMMetaSettingInfo is internal API and users cannot access
it to create their own NMSetting subclasses. But that is almost desired.
libnm is not designed, to be extensible via subclassing, nor is it
clear why that would be a useful thing to do. One day, we should remove
the NMSetting and NMSettingClass definitions from public headers. Their
only use is subclassing the types, which however does not work.
While libnm-core was linking already against nm-meta-setting.c,
nm_meta_setting_infos was unreferenced. So, this change increases
the binary size of libnm and NetworkManager (1032 bytes). Note however
that roughly the same information was previously allocated at runtime.
Not all properties that we want to handle in nmcli are actual GObject
properties. For the moment that was the case, but that's about to change.
This is a change in behavior with respect of the order in which
properties are reported. For example, print_setting_description()
now prints the property descriptions in a different order. However,
one might argue that this order makes more sense because:
- it's the same order as properties are listed in
"nm-meta-setting-desc.c". At that place, we have explict
control over the order and set it intentionally to suite
our needs best. The order of the GObject properties is
much less well defined.
- the order from "nm-meta-setting-desc.c" is used at several other
places. So, it makes sense to use the same order everywhere.
- Don't use @parent_class name. This local variable (and @object_class) is
the class instance up-cast to the pointer types of the parents. The point
here is not that it is the direct parent. The point is, that it's the
NMSettingClass type.
Also, it can only be used inconsistently, in face of NMSettingIP4Config,
who's parent type is NMSettingIPConfig. Clearly, inside
nm-setting-ip4-config.c we wouldn't want to use the "parent_class"
name. Consistently rename @parent_class to @setting_class.
- Also rename the pointer to the own class to @klass. "setting_class" is also the
wrong name for that, because the right name would be something like
"setting_6lowpan_class".
However, "klass" is preferred over the latter, because we commonly create new
GObject implementations by copying an existing one. Generic names like "klass"
and "self" inside a type implementation make that simpler.
- drop useless comments like
/* virtual functions */
/* Properties */
It's better to logically and visually structure the code, and avoid trival
remarks about that. They only end up being used inconsistently. If you
even need a stronger visual separator, then an 80 char /****/ line
should be preferred.
Properties that are backed by a GObject property are fundamentally
different.
I think it's clearer to rework the check, to first check whether
we have a param_spec, and then implement different checks.
Parsing can be complicated enough. It's simpler to just work
top-to-bottom, without calling various helper functions. This was,
you can see all the code in one place, without need to jump to
the helper function to see what it is doing.
In general, a static function that is only called once, does sometimes
not simplify but obfuscate the code.
The tests already honored the environment variable $NMTST_IFCFG_RH_UPDATE_EXPECTED
to indicate that the .cexpected files should be written by the tests.
However, in the meantime, we instead use NM_TEST_REGENERATE=1 at various
places for this purpose. Honor that flag as well.
When assuming existing connections, allow the same connection to be
activated on a different device if the connection is multi-connect
capable. Otherwise, when a connection is active on multiple devices
and NM is restarted, we assume only the first instance, and create
in-memory connections for others.
In general, a activatable connection is one that is currently not
active, or supports to be activatable multiple times according to
multi-connect setting. In addition, during autoconnect, a profile
which is marked as multi-connect=manual-multiple will not be avalable.
Hence, add an argument "for_auto_activation".
The code is mostly unused but will be used next (except for connections,
which set connection.multi-connect=multiple).
Add a new option that allows to activate a profile multiple times
(at the same time). Previoulsy, all profiles were implicitly
NM_SETTING_CONNECTION_MULTI_CONNECT_SINGLE, meaning, that activating
a profile that is already active will deactivate it first.
This will make more sense, as we also add more match-options how
profiles can be restricted to particular devices. We already have
connection.type, connection.interface-name, and (ethernet|wifi).mac-address
to restrict a profile to particular devices. For example, it is however
not possible to specify a wildcard like "eth*" to match a profile to
a set of devices by interface-name. That is another missing feature,
and once we extend the matching capabilities, it makes more sense to
activate a profile multiple times.
See also https://bugzilla.redhat.com/show_bug.cgi?id=997998, which
previously changed that a connection is restricted to a single activation
at a time. This work relaxes that again.
This only adds the new property, it is not used nor implemented yet.
https://bugzilla.redhat.com/show_bug.cgi?id=1555012
For dynamic IP methods (DHCP, IPv4LL, WWAN) the route metric is set at
activation/renewal time using the value from static configuration. To
support runtime change we need to update the dynamic configuration in
place and tell the DHCP client the new value to use for future
renewals.
https://bugzilla.redhat.com/show_bug.cgi?id=1528071
Some warnings depends on others: -Wformat-security won't work without
-Wformat. With -Wall we're confident enough that we have important
warnings enabled and in any case we're going to enable it anyway.
https://github.com/NetworkManager/NetworkManager/pull/175
g-ir-scanner uses distutils that have an unfortunate misfeature of
inheriting the compiler flags it itself was built with.
This includes the hardening flags that don't work with without
redhat-rpm-build and break with clang every full moon.
A configure check makes it clear about what went wrong in case
introspection is desired, otherwise turns it off.
(Taken from network-manager commit 678890ed0347849990787e8893122a39f95cb708)
This is a mere debugging convenience thing: e.g. if you run, but want to
check whether nm-applet or nmcli agent works fine, it's convenient that
the agent you run later gets a chance to deal with the secrets requests
first.
Is seems to do the job and is simpler that adding some more complicated
policy (e.g. introducing priorities or something).
https://github.com/NetworkManager/NetworkManager/pull/174
