Historically, the NMSetting types were in public headers. Theoretically,
that allowed users to subtype our classes. However in practice that was
impossible because they lacked required access to internal functions to
fully create an NMSetting class outside of libnm. And it also was not
useful, because you simply cannot extend libnm by subtyping a libnm
class. And supporting such a use case would be hard and limit what we can
do in libnm.
Having GObject structs in public headers also require that we don't
change it's layout. The ABI of those structs must not change, if anybody
out there was actually subclassing our GObjects.
In libnm 1.34 (commit e46d484fae ('libnm: hide NMSetting types from
public headers')) we moved the structs from headers to internal.
This would have caused a compiler error if anybody was using those
struct definitions. However, we still didn't change the ABI/layout so
that we didn't break users who relied on it (for whatever reason).
It doesn't seem there were any affected user. We waited long enough.
Change internal ABI.
No longer use g_type_class_add_private(). Instead, embed the private
structs directly (_NM_GET_PRIVATE()) or indirectly
(_NM_GET_PRIVATE_PTR()) in the object.
The main benefit is for debugging in the debugger, where we can now
easily find the private data. Previously that was so cumbersome to be
effectively impossible.
It's also the fastest possible way, since NM_SETTING_*_GET_PRIVATE()
literally resolves to "&self->_priv" (plus static asserts and
nm_assert() for type checking).
_NM_GET_PRIVATE() also propagates constness and requires that the
argument is a compatible pointer type (at compile time).
Note that g_type_class_add_private() is also deprecated in glib 2.58 and
replaced by G_ADD_PRIVATE(). For one, we still don't rely on 2.58. Also,
G_ADD_PRIVATE() is a worse solution as it supports a usecase that we
don't care for (public structs in headers). _NM_GET_PRIVATE() is still
faster, works with older glib and most importantly: is better for
debugging as you can find the private data from an object pointer.
For NMSettingIPConfig this is rather awkward, because all direct
properties require a common "klass->private_offset". This was however
the case before this change. Nothing new here. And if you ever touch
this and do something wrong, many unit tests will fail. It's almost
impossible to get wrong, albeit it can be confusing to understand.
https://gitlab.freedesktop.org/NetworkManager/NetworkManager/-/merge_requests/1773
We use clang-format for automatic formatting of our source files.
Since clang-format is actively maintained software, the actual
formatting depends on the used version of clang-format. That is
unfortunate and painful, but really unavoidable unless clang-format
would be strictly bug-compatible.
So the version that we must use is from the current Fedora release, which
is also tested by our gitlab-ci. Previously, we were using Fedora 34 with
clang-tools-extra-12.0.1-1.fc34.x86_64.
As Fedora 35 comes along, we need to update our formatting as Fedora 35
comes with version "13.0.0~rc1-1.fc35".
An alternative would be to freeze on version 12, but that has different
problems (like, it's cumbersome to rebuild clang 12 on Fedora 35 and it
would be cumbersome for our developers which are on Fedora 35 to use a
clang that they cannot easily install).
The (differently painful) solution is to reformat from time to time, as we
switch to a new Fedora (and thus clang) version.
Usually we would expect that such a reformatting brings minor changes.
But this time, the changes are huge. That is mentioned in the release
notes [1] as
Makes PointerAligment: Right working with AlignConsecutiveDeclarations. (Fixes https://llvm.org/PR27353)
[1] https://releases.llvm.org/13.0.0/tools/clang/docs/ReleaseNotes.html#clang-format
"direct" properties are the latest preferred way to implement GObject
base properties. That way, the property meta data tracks the
"direct_type" and the offset where to find the data in the struct.
That way, we can automatically
- initialize the default values
- free during finalize
- implement get_property()/set_property()
Also, the other settings operations (compare, to/from D-Bus) are
implemented more efficiently and don't need to go through
g_object_get_property()/GValue API.
String properties in libnm's NMSetting really should have NULL as a
default value. The only property that didn't, was "dcb.app-fcoe-mode".
Change the default so that it is also NULL.
Changing a default value is an API change, but in this case probably no
issue. For one, DCB is little used. But also, it's not clear who would
care and notice the change. Also, because previously verify() would reject
a NULL value as invalid. That means, there are no existing, valid profiles
that have this value set to NULL. We just make NULL the default, and
define that it means the same as "fabric".
Note that when we convert integer properties to D-Bus/GVariant, we often
omit the default value. For string properties, they are serialized as
"s" variant type. As such, NULL cannot be expressed as "s" type, so we
represent NULL by omitting the property. That makes especially sense if
the default value is also NULL. Otherwise, it's rather odd. We change
that, and we will now always express non-NULL value on D-Bus and let
NULL be encoded by omitting the property.
These functions tend to have many arguments. They are also quite som
boilerplate to implement the hundereds of properties we have, while
we want that properties have common behaviors and similarities.
Instead of repeatedly spelling out the function arguments, use a macro.
Advantages:
- the usage of a _NM_SETT_INFO_PROP_*_FCN_ARGS macro signals that this
is an implementation of a property. You can now grep for these macros
to find all implementation. That was previously rather imprecise, you
could only `git grep '\.to_dbus_fcn'` to find the uses, but not the
implementations.
As the goal is to keep properties "similar", there is a desire to
reduce the number of similar implementations and to find them.
- changing the arguments now no longer will require you to go through
all implementations. At least not, if you merely add an argument that
has a reasonable default behavior and does not require explicit
handling by most implementation.
- it's convenient to be able to patch the argument list to let the
compiler help to reason about something. For example, the
"connection_dict" argument to from_dbus_fcn() is usually unused.
If you'd like to find who uses it, rename the parameter, and
review the (few) compiler errors.
- it does save 573 LOC of boilerplate with no actual logic or useful
information. I argue, that this simplifies the code and review, by
increasing the relative amount of actually meaningful code.
Disadvantages:
- the user no longer directly sees the argument list. They would need
cscope/ctags or an IDE to jump to the macro definition and conveniently
see all arguments.
Also use _nm_nil, so that clang-format interprets this as a function
parameter list. Otherwise, it formats the function differently.
When looking at a property, it should always be clear how it is handled.
Also the "default" action should be an explicit hook.
Add _nm_setting_property_from_dbus_fcn_gprop() and set that as
from_dbus_fcn() callback to handle the "default" case which us
build around g_object_set_property().
While this adds lines of code, I think it makes the code easier to
understand. Basically, to convert a GVariant to a property, now all
properties call their from_dbus_fcn() handler, there is no special casing.
And the gprop-hook is only called for properties that are using
_nm_setting_property_from_dbus_fcn_gprop(). So, you can reason about
these two functions at separate layers.
So far, we only have NMSettingClass.compare_property() hook.
The ugliness is that this hook is per-setting, when basically
all implementations only compare one property.
It feels cleaner to have a per-property hook and call that consistently.
In step one, we give all properties (the same) compare_fcn() implementation,
which delegates to the existing NMSettingClass.compare_property().
In a second step, this will be untangled.
There is one problem with this approach: NMSettInfoPropertType grows by
one pointer size, and we have potentially many such types. That should
be addressed by unifying types in the future.
NMSetting instances either have no private data, they use
g_type_add_class_private(), or they embed the private data in the
NMSetting struct.
In all cases, we can find the private data at a fixed offset. Track that
offset in the NMSettInfoSetting meta data.
This will be useful, because properties really are stored in simple
fields, like a boolean property can be stored in a "bool" field. We will
extend the property meta data to track the offset of this property
field, but we also need to know where the offset starts.
The advantage is that we use similar macros for initializing the
static structs like
const NMSettInfoPropertType nm_sett_info_propert_type_cloned_mac_address;
and the ad-hoc locations that use NM_SETT_INFO_PROPERT_TYPE().
The former exist for property types that are used more than once.
The latter exist for convenience, where a property type is implemented
at only one place.
Also, there are few direct references to _nm_setting_property_to_dbus_fcn_gprop().
all users use NM_SETT_INFO_PROPERT_TYPE_GPROP() or
NM_SETT_INFO_PROPERT_TYPE_GPROP_INIT().
If a property can be converted to D-Bus, then always set the
to_dbus_fcn() handler. The only caller of to_dbus_fcn() is
property_to_dbus(), so this means that property_to_dbus()
has no more default implementation and always delegates to
to_dbus_fcn().
The code is easier to understand if all properties implement
to_dbus_fcn() the same way.
Also, there is supposed to be a split between NMSettInfoProperty (info about
the property) and NMSettInfoPropertType (the type). The idea is that
each property (obviously) requires its distinct NMSettInfoProperty, but
they can share a common type implementation.
With NMSettInfoPropertType.gprop_to_dbus_fcn that is often violated because
many properties that implement NMSettInfoPropertType.gprop_to_dbus_fcn
require a special type implementation. As such, gprop_to_dbus_fcn should
be part of the property info and not the property type. The first step towards
that is unifying all properties to use to_dbus_fcn().
When subclassing a GObject type, the class and object structs
must be available and defined in the header.
For libnm, and in particular for NMSetting classes, we don't want
users to subclass NMSetting. It also doesn't work, because libnm
has internal code that is necessary to hook up the NMSetting class.
You cannot define your own type and make it work together with
libnm.
Having the structs in public headers limits what we can do with them.
For example, we could embed the private data directly in the structures
and avoid the additional indirection.
This is an API break, but for something that most likely nobody cares
about. Or better, nobody should care about. API is not what is
accidentally defined in a header, API was the library provides to
meaningfully use. Subclassing these types is not meaningful and was
only accidentally possible so far.
Only hide the structs for now. More cleanup is possible later. We shall
however aim to keep the padding and struct layout to not also break ABI.
(cherry picked from commit e46d484fae)
"libnm-core/" is rather complicated. It provides a static library that
is linked into libnm.so and NetworkManager. It also contains public
headers (like "nm-setting.h") which are part of public libnm API.
Then we have helper libraries ("libnm-core/nm-libnm-core-*/") which
only rely on public API of libnm-core, but are themself static
libraries that can be used by anybody who uses libnm-core. And
"libnm-core/nm-libnm-core-intern" is used by libnm-core itself.
Move "libnm-core/" to "src/". But also split it in different
directories so that they have a clearer purpose.
The goal is to have a flat directory hierarchy. The "src/libnm-core*/"
directories correspond to the different modules (static libraries and set
of headers that we have). We have different kinds of such modules because
of how we combine various code together. The directory layout now reflects
this.
2021-02-18 19:46:51 +01:00
Renamed from libnm-core/nm-setting-dcb.c (Browse further)
